Studded Tires and Highway Safety Feasibility of Determining Indirect Benefits

176

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM REPORT 176 STUDDED TIRES AND HIGHWAY SAFETY FEASIBILITY OF DETERMINING INDIRECT BENEFITS

TRANSPORTATION RESEARCH BOARD NATIONAL RESEARCH COUNCIL TRANSPORTATION RESEARCH BOARD 1977 Officers ROBERT N. HUNTER, Chairn,a,, SCHEFFER LANG, Vice Chairman W. N CAREY JR., Executive Director

Executive Committee HENRIK F. STAFSETH, Executive Director, An,erica,, Ass,,, of State Highway and Tratisportatio,, Officials (ex officio) WILLIAM M. COX, Federal Highway Admi,,istrator, U.S. Department of Tra,,sportatio,, (ex officio) RICHARD S. PAGE, Urban Mass Transportation Admtnistrator, U.S. Department of Transportation (ex officio) JOHN M. SULLIVAN, Federal Railroad Administrator, U.S. Department of Transportation (ex officio) HARVEY BROOKS, Chairman, Com,nission on Sociotechnical Syste,,,s, National Research Council (ex officio) MILTON PIKARSKY, Chairman of the Board, Chicago Regional Transportatio,, Authority (ex officio, Past Chairman 1975) HAROLD L. MICHAEL, School of Civil Engineering, Purdue University (ex officio, Past Chairman 1976) WARREN F. ALBERTS, Vice President (Syste,ns Operations Services), U,,ited Airli,,es GEORGE H. ANDREWS, Vice President (Transportation Marketing), Sverdrup and Parcel GRANT BASTIAN, State Highway Engineer, Nevada Department of Highways KURT W. BAUER, Executive Director, Southeastern Wisconsin Regional Pla,,ni,,g Con,m,ssio,, MANUEL CARBALLO, Madison, Wisconsin L. DEBERRY, Engineer-Director, Texas State Department of Highways and Public Transportation LOUIS J. GAMBACCINI, Vice President and General Manager, Port Authority Trans-Hudso,, Corporation HOWARD L. GAUTHIER, Professor of Geography, Ohio State University FRANK C. HERRINGER, General Manager, San Francisco Bay Area Rapid Transit District ARTHUR J. HOLLAND, Mayor, City of Trenton, N.J. ANN R. HULL, Speaker Pro Tern, Maryland House of Delegates ROBERT N. HUNTER, Chief Engineer, Missouri State Highway Department PETER G. KOLTNOW, President, Highway Users Federation for Safety and Mobility THOMAS I. LAMPHIER, President, Transportatiot, Division, Burlington Northern, Inc. A. SCHEFFER LANG, Assistant to the President, Association of American Railroads DANIEL McFADDEN, Professor of Economics, University of California ROBERT S. MICHAEL, Director of Aviation, City and County of Denver, Colorado THOMAS D. MORELAND, Commissioner, Georgia Department of Transportation GEORGE E. FAKE, Vice President, Xerox Corp.; Manager, Xerox Palo Alto Research Center DOUGLAS N. SCHNEIDER, JR., Director, District of Columbia Department of Transportation WILLIAM K. SMITH, Vice President (Transportation), General Mills

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM Transportation Research Board Executive Committee Subco,nmittee for the NCHRP ROBERT N. HUNTER, Missouri State Highway Department (Chairman) A. SCHEFFER LANG, Association of American Railroads HENRIK E. STAFSETH, Amer. Ass,,, of State Hwy. and Transp. Officials WILLIAM M. COX, U.S. Department of Transportation HARVEY BROOKS, National Research Council HAROLD L. MICHAEL, Purdue University W. N. CAREY, JR., Transportation Research Board

General Field of Design Area of Pavements Project Panel C1-13(2) W. S. EKERN, Minnesota Department of Highways (Chairman) PHILLIP S. MANCINI, Retired Raymond, Parish & Pine, Inc. MALCOLM D. ARMSTRONG, Ministry of Transport, Montreal, Canada WILLIAM R. McGRATH, WILLIAM F. BAUCH, Federal Highway Administration JACK RECHT, National Safety Council CECIL BRENNER, National Highway Traffic Safety Administration E. A. WHITEHURST, Ohio State University BRADFORD M. CRITFENDEN, National Hwy, Traffic Safety Admit,. GLENN G. BALMER, Federal Highway Administratio,, KARL H. DUNN, Wisconsin Department of Transportation K. B. JOHNS, Transportation Research Board DAVID MAHONE, Virginia Highway Research Council

Program Staff KRIEGER W. HENDERSON, JR., Program Director HARRY A. SMITH, Projects Engineer DAVID K. WITHEFORD, Assistant Program Director ROBERT E. SPICHER, Projects Engineer LOUIS M. MACGREGOR, Administrative Engineer HERBERT P. ORLAND, Editor R. IAN KINGHAM, Projects Engineer PATRICIA A. PETERS, Associate Editor ROBERT J. REILLY, Projects Engineer EDYTHE T. CRUMP, Assistant Editor NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM REPORT 1 76

STUDDED TIRES AND HIGHWAY SAFETY FEASIBILITY OF DETERMINING INDIRECT BENEFITS

J. S. CRESWELL, D. F. DUNLAP AND J. A. GREEN HIGHWAY SAFETY RESEARCH INSTITUTE THE UNIVERSITY OF MICHIGAN

RESEARCH SPONSORED BY THE AMERICAN ASSOCIATION OF STATE HIGHWAY AND TRANSPORTATION OFFICIALS IN COOPERATION WITH THE FEDERAL HIGHWAY ADMINISTRATION

AREAS OF INTEREST: TRANSPORTATION ADMINISTRATION PAVEMENT DESIGN PAVEMENT PERFORMANCE MAINTENANCE GENERAL HIGHWAY SAFETY

TRANSPORTATION RESEARCH BOARD NATIONAL RESEARCH COUNCIL WASHINGTON, D.C. 1977 NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM NCHRP Report 176

Systematic, welt-designed research provides the most ef- Project 1-13(2) FY'72 fective approach to the solution of many problems facing ISBN 0-309-02543-5 highway administrators and engineers. Often, highway L. C. Catalog Card No. 77-85123 problems are of local interest and can best be studied by highway departments individually or in cooperation with Price: $4.00 their state universities and others. However, the accelerat- ing growth of highway transportation develops increasingly complex problems of wide interest to highway authorities. These problems are best studied through a coordinated program of cooperative research. In recognition of these needs, the highway administrators Notice of the American Association of State Highway and Trans- The project that is the subject of this report was a part of the portation Officials initiated in 1962 an objective national National Cooperative Highway Research Program conducted by the highway research program employing modern scientific Transportation Research Board with the approvat of the Governing techniques. This program is supported on a continuing Board of the National Research Council, acting in behalf of the National Academy of Sciences. Such approval reflects the Governing basis by funds from participating member states of the Board's judgment that the program concerned is of nationat impor- Association and it receives the full cooperation and support tance and appropriate with respect to both the purposes and re- sources of the National Research Council. of the Federal Highway Administration, United States The members of the technical committee selected to monitor this Department of Transportation. project and to review this report were chosen for recognized The Transportation Research Board of the National Re- scholarly competence and with due consideration for the balance of disciplines appropriate to the project. The opinions and con- search Council was requested by the Association to admin- clusions expressed or implied are those of the research agency that ister the research program because of the Board's recog- performed the research, and, while they have been accepted as appropriate by the technical committee, they are not necessarily those nized objectivity and understanding of modern research of the Transportation Research Board, the National Research Coun- practices. The Board is uniquely suited for this purpose cil, the National Academy of Sciences, or the program sponsors. as: it maintains an extensive committee structure from Each report is reviewed and processed according 10 procedures established and monitored by the Report Review Committee of the which authorities on any highway transportation subject National Academy of Sciences. Distribution of the report is ap- may be drawn; it possesses avenues of communications and proved by the President of the Academy upon satisfactory comple- cooperation with federal, state, and local governmental tion of the review process, The National Research Council is the principat operating agency of agencies, universities, and industry; its relationship to its the National Academy of Sciences and the National Academy of parent organization, the National Academy of Sciences, a Engineering, serving government and other organizations. The private, nonprofit institution, is an insurance of objectivity; Transportation Research Board evotved from the 54-year-old High- way Research Board. The TRB incorporates all former FIRB it maintains a full-time research correlation staff of special- activities but also performs additional functions under a broader ists in highway transportation matters to bring the findings scope involving all modes of transportation and the interactions of of research directly to those who, are in a position to use transportation with society. them. The program is developed on the basis of research needs identified by chief administrators of the highway and transportation departments and by committees of AASHTO. Each year, specific areas of research needs to be included in the program are proposed to the Academy and the Board by the American Association of State Highway and Trans- portation Officials. Research projects to fulfill these needs are defined by the Board, and qualified research agencies are selected from those that have submitted proposals. Ad- ministration and surveillance of research contracts are Published reports of the responsibilities of the Academy and its Transportation NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM Research Board. The needs for highway research are many, and the National are available from: Cooperative Highway Research Program can make signifi- Transportation Research Board cant contributions to the solution of highway transportation National Academy of Sciences problems of mutual concern to many responsible groups. 2101 Constitution Avenue, N.W. The program, however, is intended to complement rather Washington, D.C. 20418 than to substitute for or duplicate other highway research programs. Printed in the United States of America.

FORE\AIORD This report synthesizes available knowledge relevant to the adverse safety effects of pavement wear and pavement marking wear attributable to studded tires, By Stafl assesses qualitalively the relative importance of the effects that were judged likely Transportation to be of consequence, and outlines research that could assist in quantifying these Research Board effects. The information that is provided should be useful to individuals and agencies who may plan to undertake research directed at a quantification of these factors.

Not uncommonly, new products that are intrdduced to aid in the solution of one problem create others. Eventually, a determination must be made as to whether the positive results outweigh the negative sufficiently to encourage continued use of the products. So it is with studded tires. Since their introduction in the early 1960's, they have received wide acceptance as a convenience item. The degree of safety that they add has been the subject of considerable controversy, and is still not delineated to everyone's satisfaction. Few now question their potential for causing pavement wear. The present study is a first step into the relatively unexplored area of possible adverse side effects that studded tires may have on safety as a result of the pavement wear and pavement marking wear that is attributed to them. Numerous indirect effects that might result from excess pavement wear were hypothesized and examined by the University of Michigan researchers. Examples are tire hydroplaning resulting from water entrapment in the channels cut in the wheelpaths by studded tires, increased splash and spray resulting from the same condition, reduced skid resistance because of the abrading effects of studded tires, and the hazards of more frequent maintenance necessary to overcome the effects of studded tires. Conceptual models were established to assist in studying the hypothesized variables of interest. Available information did not permit quantification of any of the effects. Broad experimental plans that can lead to the quantification of what seem to be the most important of the indirect effects were developed and the extent and cost of the work that would be necessary was assessed. It appears that the task of obtaining truly definitive information on the side effects of studded tires, that could tip the balance in favor of or against their continued use, will be one of formidable proportions.

CONTENTS

1 SUMMARY

PART I

2 CHAPTER ONE Introduction Problem and Objectives Research Plan

4 CHAPTER TWO Research Findings Hazardous Effects Tire Hydroplaning and Wet Skid Pavement Maintenance Hazard Splash and Spray Vehicle Lateral Placement Shifting Vehicle Transverse Forces and Steering Effects Driver Fatigue Resulting from Noise and Vibration Ejected Studs Thrown from High-Speed Vehicles Vehicle Component Degradation

8 CHAPTER THREE Suggested Research Accident Causation Mechanisms In-Service Mechanisms Identification Accident Data Analysis

PART II

10 APPENDIX A Synthesis of Research Results

24 APPENDIX B In-Service Prediction of Hydroplaning and Wet Skid

27 APPENDIX C Proposed Research Plan

36 APPENDIX D Bibliography ACKNOWLEDGMENTS The research reported herein was conducted as NCHRP Project 1-13(2) by the Highway Safety Research Institute, The Univer- sity of Michigan, Ann Arbor. John A. Green, Research Physi- cist, acted as project director, with Duane F. Dunlap, Research Engineer, serving as co-principal investigator responsible for the engineering aspects of the study, and Jay S. Creswell, Research Associate, serving as co-principal investigator responsible for conceptual-model development and a design for future research. STUDDED TIRES AND HIGHWAY SAFETY FEASIBILITY OF DETERMINING INDIRECT EFFECTS

SUMMARY Wear of pavement and pavement marking by studded tires is suspected as the cause of several effects that result in decreased highway safety. In the order of decreasing degree of hazard, the most important effects are (a) tire hydroplaning and wet skid and (b) pavement maintenance hazard. Other effects, discussed in more detail in the report, are believed to be of less potential hazard than those cited. In the case of hydroplaning, studded tire wear was found to be both beneficial and detrimental to safety. In some cases, a coarsening of the surface produces enhanced skid-resistance properties. In other cases, studded tire wear has a smoothing effect on the surface, which reduces skid number. If projections of road maintenance activity resulting from studded tire wear hold true, construction sites may be a significant source of accidents. The decision to repave or recondition a road surface is determined at least partially by the known or suspected relationship between road damage and accidents. Thus extensive repairs partially assume a knowledge of the known accident causation factors of damaged pavement. - Some possible studded tire effects, such as the ejection of studs from high- speed vehicles or the degradation of vehicle components, were judged to have little relationship to accidents. Modeling activities have suggested that the relationship between pavement wear effects and accident rate may be difficult to define in an experimental context. A suggested approach plan involves four major investigative phases: Definition of the extent of road damage. Examination of available accident data. Observations of traffic flow patterns on lightly and heavily damaged road sections. Collection of supplemental information during police investigations. Studies indicate that accident information corresponding to a year's experience on the New York Thruway or the Ohio Turnpike will be necessary to produce results that are statistically significant. CHAPTER ONE

INTRODUCTION

PROBLEM AND OBJECTIVES to date, extremes of wear have been confined, fortunately, to limited areas where circumstances have been favorable to Following their introduction into North America in the wear development. Care must be exercised in generalizing mid-1960's, studded tires have received rapid motorist ac- such effects to the entire highway system. ceptance. Their adoption and continued use, however, has Under normal practice, pavement surfaces are sloped been a source of much controversy in the highway com- transversely to provide for quick runoff of water from rain munity. Advocates of studded tires argue for their con- and melting snow. The presence of stud-produced troughs tinued use, claiming that they improve safety under winter that inhibit drainage fosters hydroplaning and wet skidding. driving conditions. Opponents of studded tires maintain Entrapment of water by troughing also affects safety by that they accelerate highway pavement wear substantially, increasing "wheel splash" (the ejection of water and slush thereby adding to pavement maintenance costs. to the side or rear of moving vehicles) and 'wheel spray" Both sides possess experimental evidence to support their (water entrained in the turbulent wake of a moving vehi- views. Tests by Smith, dough, and others (205-208) indi- cle). Severe troughing also influences the lateral placement cate that vehicles equipped with studded tires have better of vehicles and the steering task. tractive performance than either conventional tires or non- The change in skid resistance of surfaces brought about studded snow tires for certain winter driving conditions. by the abrasive action of studs is another indirect safety Studded tires' greatest relative merit is on glare ice at 32 F. influence that can be hypothesized to be either positive or But climatological studies indicate that such optimal con- negative depending on pavement wear characteristics. In- ditions occur infrequently. On the other hand, various creased accident potential resulting from intensified main- pavement wear experiments and observations of actual road tenance operations required to repair stud damage is an- surfaces indicate that studded tire use accelerates pavement other. All of the foregoing possible safety influences, and wear (104, 153-155, 171). But studies of relative accident others, have been examined in this research. involvement have not shown any clear-cut degradation in highway safety as a result of this wear. Given the evidence RESEARCH PLAN to date, opponents of studded tire use argue that the demon- strable costs exceed the unproved benefits; consequently, As indicated earlier, the present state of knowledge con- the use of studded tires ought to be banned, as Minnesota cerning the safety effects of studded tires is such that the and Ontario have done. highway community can benefit from a collection and syn- Ultimately, the fate of studded tires should be resolved thesis of the available research results. This study was on the basis of the relative costs and benefits of their use. designed to fill this need. At present, not enough is known about the problem to One of the first tasks in the program was the preparation achieve such a resolution. Efforts are continuing to de- of a bibliography of published material relative to studded termine the winter driving safety and convenience benefits tire pavement wear, which is given in Appendix D. The of studded tires and to refine the estimates of probable citations are grouped according to subject area and are pavement damage. What remain to be determined are the listed alphabetically therein. indirect effects on highway safety that can be assigned to An intensive review of the literature resulted in a defini- the use of studded tires. tion of the important categories of safety effects and de- The principal indirect effects of studded tires on high- lineation of those particular areas where data were avail- way safety result from the accelerated wear of pavement able. A detailed synthesis of the literature review is pre- surfaces and obliteration of pavement markings. Pavement sented in Appendix A. The material obtained from this wear is a function of the paving material, the channelizing review was used to develop accident causal chain models— character of the roadway, the exposure mileage, and the first on a global basis, and then for each separate effect. speed distribution of vehicles equipped with studded tires. The models, coupled with the available information, were Wear is most pronounced where traffic channelization, large used to identify areas where additional knowledge was vehicle accelerations, and high exposure occur simulta- required and, thus, were used to develop the follow-on neously. The fact that such troughing can be caused by research recommendations presented in Appendix C. A studded tires has been amply demonstrated (5, 29, 95, 164, cursory presentation of the model, which was central to /66, 183). Toll booth lanes and freeway entrance and exit the strudturing of the problem, is given here, and a some- ramps are examples of particularly high-wear areas. Actual what fuller description follows in Appendix A. ruts or troughs I in. and more in depth have been observed The model was used to guide initial efforts in classifying at these locations (/03, 178). The troughs have a shape the literature and organizing further research efforts. Once similar to a normal frequency distribution curve, and are the literature review had been completed, the model was characteristically three or more times wider than a tire. Yet revised to reflect a fuller understanding of the process. Then, in the absence of hard data on many of the problems, The model, shown in Figure 1, has three principal seg- the model was used to provide a basis for informed judg- ments: ( 1 ) preconditions, (2) active factors, and (3) main ment to select those areas where more serious hazards were effects. The preconditions segment details situations in likely to occur. The judgment was that the more impor- which substantial studded tire damage is likely to arise, tant effects would be found for wet-weather, loss-of-control the active factors section indicates the effects of the dam- crashes on high-traffic-volume interstate-type highways. age on the accident causation process, and the main effects portion shows the combination of preconditions and active factors likely to have the most substantial effect on accidents. Total Traffic Volume

I Stud Other Design Materials I Equipped Vehicles I Vehicles PRECONDITIONS Climate

Pavement 4 I Pavement Maintenance I Wear

I ......

Pavement - Changes in Changes in Marking Changes in I Lateral Riding Char-a Frictional Profile acteristics!. Properties (Rutting)

ACTIVE FACTORS Traffic Flow Water Accumulation Characteristics (Depth and Pattern)

V Driver Incidence onses Frequency of of Hydrop laning Conflict Situations Vehicle Properties Frequency + ofLossof .4 I Dry Control Weather Crashes

Wet 1 MAIN EFFECTS Weather Crashes r...... Figure 1. General model of stud-induced damage and possible consequences. CHAPTER TWO

RESEARCH FINDINGS

HAZARDOUS EFFECTS Wet skid, which usually involves partial hydroplaning, was found to be primarily influenced by surface texture Wear of pavement and pavement marking by studded tires characteristics. A gritty microtexture combined with an is suspected of causing several effects that result in degraded open macrotexture gives the best skid resistance, both in traffic safety. The effects found to be most hazardous in terms of low-speed performance and insensitivity to speed. the Current program are listed in the order of estimated Maintaining good friction characteristics at higher speeds decreasing hazard: is most influenced by an open macrotexture. Studded tire Tire hydroplaning and wet skid. wear was found to be both beneficial to and detrimental to Pavement maintenance hazard (the result of pave- maintaining good surface skid resistance. In some cases ment surface and pavement marking restoration). (e.g., topeka or asphaltic concrete) studded tire wear in the Splash and spray. winter coarsens the surface and produces better skid- Vehicle lateral placement shifting. resistance properties. In other cases, studded tire wear has Vehicle transverse forces and steering effects. a smoothing effect, particularly when light vehicles are Driver fatigue resulting from noise and vibration. involved and the skid number is reduced. Appendix A, Ejected studs thrown from high-speed vehicles. Section A.1.5, gives a detailed discussion of this effect. Vehicle component degradation. A nomograph that illustrates a means, which further study may show to have very practical application, aids in A detailed discussion of each effect is presented, with an assessing the relative sensitivities of the variables that in- assessment of its probable importance to highway safety. fluence dynamc hydroplaning and wet skid and in deter- Ranking is based on the assessment of the relative like- mining the vehicle speeds at which they might take place. lihood of the simultaneous presence of both preconditions The nomograph is shown in Figure 2. Details regarding its and active factors for the hypothesized effect. These assess- development are presented in Appendix B. ments come from the best available data or, lacking those, engineering judgment guided by the conceptual model pre- PAVEMENT MAINTENANCE HAZARD sented earlier. For example, tire hydroplaning and wet skid is rated the most hazardous because the precondition Past experience shows that maintenance activities repre- (water-filled troughs) is likely to become common if sent a significant hazard (87, 88). Construction-associated studded tire use continues. In addition, traffic situations accidents are a small fraction, 1 percent to 3 percent, of all (active factors) in which pavement friction is critical to the crashes. Yet, this proportion is high compared to mileage accident generation process are quite frequent. Conversely, under construction. Also, based on the ratio of fatal to vehicle component degradation is ranked lowest because nonfatal involvements, these crashes appear to be some- items such as steering and suspension degradation would be what more severe than average. Historical data for Texas infrequently increased, if at all, by driving on stud-damaged are listed in Table 1. pavements and because situations where performance of Several highway organizations have projected vast ex- these components is critical rarely occur. penditures for repairing roads damaged by studded tires (9, 12, 83, 103). Even if only a small fraction of these TIRE HYDROPLANING AND WET SKID projections become true, repair and maintenance activities will increase markedly. The associated accident potential Tire hydroplaning and wet skid result from a complex set can rise significantly as a consequence. Because no known of tire, vehicle, and road surface factors. Tire factors in- studies deal specifically with road surface repair activities, clude inflation pressure, tread pattern, tread width, tread the increase cannot be exactly predicted. Nevertheless, be- depth, carcass construction, contact path length, deflection, cause such repairs are likely to be in dangerous traffic con- and material; for the vehicle, factors are wheel load and ditions, pavement maintenance is rated the second most suspension system; and surface factors include texture hazardous consequence of studded tire damage. depth, cross slope, drainage path length, rainfall intensity, microtexture (friction-producing), and macrotexture (drain- SPLASH AND SPRAY age path outlets). Dynamic hydroplaning, or complete lift-off, is probably a Ruts caused by studded tires can be expected to increase rare event under ordinary conditions of surface drainage, the prevalence and intensity of visibility interference due surface texture, tire tread, and highway speed (17). The to splash and spray. Although annoying and bothersome, production of pavement wheel ruts by studded tires (5, 29, splash and spray effects prior to studded tire wear expe- 95, 164, 166, 183) can, on the other hand, be suspected of rience have not been identified as a significant accident trapping sufficient amounts and depths of water to increase producer (less than 0.2 percent of reported accidents) significantly the probability of dynamic hydroplaning. (92). Most such accidents are the result of a panic brak- 0 A 96 B ho A'

-2 4 0.05

3 0.2 F-0 2 I 0.10

I- 0.1

0 -BFO.ls ' . ci- I; -1 'J.00.2 .c ._ -2 L0 10 43' tI 0 r0.3 (9 O H 0.25 k0.4 24 -0.30 /

(Find A Value on A' Scale) k- 0.35

Tire Lift Coefficient (ce) 30 2 A. Conventional SM1PLE SOLUTION smooth ioo long, rib tread (1/4 in 26 S = 48 closed tread (1/4 in) 18 I = 0.8 in./hr open tread (1/4 iii) 15 S = 20 ft B ' / B. Radial - T = 0.10 in. smooth R = 0.20 in. 60 C closed tread H 25 01 50 25 K 0.65 Ho open tread 16 N Solution = 24 mph aSaa Pavement Surface Factor (K) C 60 40 A Smooth F0.02 6 0.15 B' SO -B Fine textured, rounded 0.38 E Fine textured, gritty 1.00 70 30 Coarse textured, rounded 0.65 & Coarse textured, gritty 2.90 0 I CoCA0 005 ' 1 90 r 20 BA0 720 o 0a 0.10 NOTE: FOR DEMONSTRATION ONLY 0:20 90 [030 10 F 0.50 (Find B Value on B' scale) Figure 2. Hydroplaning or wet.skid prediction notnograph. ing maneuver on the part of the driver during an obscured- TABLE I windshield experience. TEXAS' DATA ON RURAL MOTOR VEHICLE Water depth influences splash intensity but not spray. ACCIDENTS AT AREAS UNDER CONSTRUCTION Studded tire pavement ruts can, therefore, increase the number of splash-induced accidents. No applicable studies ACCJDENT5 have been made and, in their absence, splash and spray YEAR rATAL NONFATAL accidents are considered to be a less serious result of 1960 60 1,664 studded tire wear than the aforementioned effects. 1961 49 1,598 1962 50 1,450 1963 56 1,517 VEHICLE LATERAL PLACEMENT SHIFTING 1964 53 2,131 1965 30 2,033 Vehicle lateral placement shifting can occur as the result 1966 53 2,341 of pavement marking obliteration and wheelpath wear. The 1967 54 2,650 former results in a loss of delineation lines that the driver uses in guiding his vehicle in lane holding. The latter produces several adverse effects that can cause the driver to steer to one side of the worn paths. Among these effects are (a) retained water in the path ruts that can cause hydroplaning, cause splash and spray problems, or freeze; (b) side forces resulting from traveling in a studded tire rut roughened pavement that can cause annoying vehicle in- having sloped sides is expected to be lower than that ex- terior noise and vibration levels; and (c) adverse steering perienced with vertical ridges. How much lower depends effects from the tendency of tires to climb the sides of the on the slope and depth of the rut. ruts. - The ridge existing at the pavement edge is considered by Although all of these factors could produce lateral place- some to be a main contributor to loss of control in ran-off- ment shifting and thus contribute to •increased accident the-road accidents. It may be, therefore, that stud-induced rates, none (with the possible exception of pavement mark- ruts having severe depth and high side slopes have a similar ing obliteration) has been proven to do so. Road edge effect in producing adverse transverse forces and steering pavement markings have been found to contribute signifi- effects within a traffic lane. An answer to this question cantly to reduced accident rates (105). The same, so mr requires a more thorough examination of tire rut climbing as is known, has not been shown for intermediate lane properties. markings on multilane highways. A study to determine the influence of wheelpath wear on DRIVER FATIGUE RESULTING FROM NOISE AND VIBRATION lateral placement shifting has produced negative results Noise and vibration effects on driver fatigue are compli- (/99). The method of taking data is open to question, cated by the fact that such effects are measured in terms however, in that only the lateral vehicle placement at a that are largely subjective. point location was measured, rather than that of the ve- Interior noise levels in the average passenger car are lust hicle path over a finite interval. It appears, therefore, that at the dividing line at which ordinary conversation can be additional studies of the effects of intermediate lane mark- carried on without great effort (131). So far as fatigue ings on lateral placement, as well as more refined investiga- discomfort is concerned, interior noise levels are within the tions of placement shifting patterns, are in order. so-called noisy comfort criterion but above the quiet comfort criterion. (The noisy comfort criterion is defined to be a maximum permissible sound pressure level below VEHICLE TRANSVERSE FORCES AND STEERING EFFECTS which people are comfortable if such people expect a noisy Adverse side forces and steering effects can occur as the environment. The quiet comfort criterion is defined in a result of vehicle tires interacting with stud-produced wheel like manner.) ruts. Tires, in general, have a tendency to climb the side Road surface effects can increase over-all vehicle noise of a ridge, or sloped surface. Although published evidence by as much as 10 dB(A) (about 15 percent) from smooth tarmac to a level cobbled surface (126). Wet conditions is conflicting, the tendency seems to be less for tires of can further increase the noise level from 3 to 15 db(A). radial construction. In the worst case involving a wet, rough surface, auto- Two phenomena have been identified by Marshall, et al. mobile interior noise levels can exceed 90 db(A) over a (109): (a) a tire running up against a low vertical ridge narrow frequency band- Above the 90 db(A) level, a slight will tend to "hold off" until sufficient side force is pro- decrease in human performance has been recorded in carry- duced to cause the tire to climb; and (b) once climbing is ing out complex tasks (95). Although the results are highly initiated, a side force tending to reinforce climbing quickly subjective, it is possible that noise levels resulting from driv- builds up. A ridge as low as 3/s in. can cause a rapidly ing over a wet, stud-roughened surface can produce long- changing vehicle side force exceeding 100 lb. Data of this term driver fatigue. type, however, are very limited and tend to be more sub- Vehicle vibrations resulting from studded tire damaged jective than objective. In any case, the comparative level of roadways are considered in the general subject area of ye- hide ride. Human response to whole-body vibration is gen- If the tire is rolling near zero slip conditions (ordinary erally used as the measure for ride comfort. Thus, the travel), the highest stud kinetic occurs when the stud is vibration path which is most pertinent to ride considera- ejected at the top of the tire (4mV2). The kinetic energy tions includes, in order, the road, the tires, the vehicle sus- is zero for a stud ejected in the contact patch directly below pension, the vehicle body, and the vehicle seat. However, the spindle. Practically, the only ejection conditions of each of these produces an attenuating effect on the road interest are those that involve a stud being thrown back- surface forcing function such that the vibration wave form ward in a low trajectory which just clears the rear over- felt by the passenger differs substantially from that occur- hang of the vehicle (143). Under such conditions the stud ring in the road. Further, individual differences in the would have little forward kinetic energy of its own and aforementioned vehicle components can produce an order would be a hazard only as the result of the velocity of a of magnitude difference in passenger-sensed effects in a following vehicle. The danger is, therefore, similar to that repeat traverse of the same road surface. existing from an ejected stone of similar weight. Comparison of seat-monitored vibration data (testing was in the early years of studded tire use in Michigan a rash done on an old runway, which was reasonably level with of reports were made that police cars were being shot at broken surface areas fairly evenly distributed) with pro- during freeway patrol. Evidence seems to suggest that most posed fatigue-decreased proficiency' limits for vertical vi- of the events were the result of ejected studs. Such reports bration, as established by the International Organization of have become relatively uncommon in more recent years. Standardization (ISO), suggests that road-induced vehicle This suggests that (a) the cause of the problem has been vibration levels are just below those that induce fatigue even identified, or (b) stud-fastening methods have improved. for periods as long as eight hours (128). Roughened road surfaces (i.e., black top versus Belgian blocks) have been VEHICLE COMPONENT DEGRADATION found to increase vibration acceleration levels by as much as a factor of four, however (126). On a very rough road Component degradation mechanisms are a function of (Belgian blocks), resulting vibrations are probably not tol- aging and loading cycles. Stud-damaged road surfaces may erable for more than about an hour without some impair- affect the latter but have no effect on aging processes. The ment in performance. The worst examples of stud-damaged major components •affected are those which are a part of roads are somewhere between black top and Belgian block the steering and suspension systems. With proper lubrica- surfaces, and, according to proposed ISO standards, are tion, however, the level of increased wear resulting from estimated to cause some vibration fatigue performance im- driving over stud-damaged pavements is considered to be pairment after about four hours of driving. These conclu- minor. sions are obviously quite tentative, however, because (a) no Road asperities caused by studded tires, which could specific stud-damaged pavement test data are available, and contribute to suspension-component cycling, are of two (b) the ISO standards are only in the "proposed" stage, types: with much of the criteria data being somewhat subjective and subject to interpretation. Exposed aggregate. Four hours of continuous driving is generally longer than Longitudinal ruts. most people drive without resting, Fatigue from sources other than vibration is of additional significance under such Exposed aggregate asperities are of an order of magni- conditions. Tentatively, therefore, it must be concluded that tude in size which is easily engulfed by the tire tread. vibration-induced fatigue is not a major factor in accident Suspension component cycling is, therefore, not involved. causation resulting from studded tire pavement wear. Noise Longitudinal ruts, especially those with steep edges, can levels from stud-worn pavements under wet conditions are produce significant steering system inputs. Traveling from considered to be slightly more fatigue producing. Noise- side to side within a rut could, therefore, increase the induced fatigue must be considered a minor accident- cycling times of steering system components. The forces causative factor compared to other effects discussed earlier, and cycling which are necessary to produce long-term however. degradation would undoubtedly be sensed by the driver as annoying, however. The driver, in turn, could be expected to steer out and travel to one side of the ruts, thus eliminat- EJECTED STUDS THROWN FROM HIGH-SPEED VEHICLES ing the disturbance. As noted earlier, no tendency for steer- The hazard resulting from studs being ejected from high- ing out of ruts has been identified. Vehicle component wear speed vehicles was found to be similar to that of an ejected as the result of stud-damaged pavements is, therefore, con- stone of the same weight. It was reported from Sweden sidered to be of minor consequence. No conclusive assess- that for a stud to be ejected by centrifugal force, a vehicle ment can be given, however, due to the general lack of speed on the order of 500 mph would be required (141). pertinent data on in-service component wear characteristics. In general, stud ejection is normally caused by the sliding Specifically, no data have been found that show the effect action between the tire and road surface. Ejected studs have of pavement surface properties on steering and suspension been found to have little kinetic energy and are often lost system wear rates. during low-speed travel in low gear. Even if component wear was found to be substantial, however, the effect on vehicle safety and subsequent acci- Shock absorbers 50 percent reduction in damping. dent causation is problematic. Recent studies have shown Ball joints No noticeable effects with artificial free- that many vehicle components must be worn considerably play of 0.1 in. radian and 0.6 in. axial. before any effect on vehicle performance is apparent (145, Steering play 2.5 in. of steering wheel freeplay. 147). Required degradation to produce "noticeable" effects for various steering and suspension components are listed These values represent substantial degradation in compo- as follows: nent performance.

CHAPTER THREE

SUGGESTED RESEARCH

As a result of the literature review and modeling activi- Vehicle transverse forces and steering effects—Ex- ties, it is evident that a variety of new information must be perimental research to determine tire nibbling characteris- gathered to explicate further the relationship between pave- tics in the presence of studded tire-like wear ruts. Both ment wear caused by studded tires and accident causation. moment and side force characteristics should be examined Appendix C is the detailed presentation of a two-phase re- for representative tire types and rut profiles. search plan. A summary of those results is given. Driver fatigue resulting from noise and vibration— Follow-on research recommendations concern three sepa- Experimental research to measure noise and vibration levels rate areas: inside a cross-section of vehicles, at various speeds, over typical studded tire damaged pavements. I. Accident causation mechanisms. Ejected studs thrown from high-speed vehicles—No In-service mechanism identification. research recommended. Accident data analysis. Vehicle component degradation—No research recommended. ACCIDENT CAUSATION MECHANISMS IN-SERVICE MECHANISMS IDENTIFICATION Accident causation mechanisms are those effects that are suspected of contributing to degraded traffic safety. Each The recommended research described here deals with mea- mechanism is listed in order, along with a brief description suring studded tire pavement and marking wear patterns of the research needed to close identifiable knowledge gaps. and resulting traffic flow patterns. The object is to identify With knowledge gained by means of the recommended re- accident causation mechanisms in actual in-service situa- search, more definitive answers can be produced relative to tions. It is suggested that one of the major turnpike sys- the actual hazard of each suspected mechanism. tems, such as the New York Thruway or the Ohio Turn- pike, be used as the location for the investigations because Tire hydroplaning and wet skid—Experimental re- these types of roads have excellent traffic data. Their semi- search (a) to establish hydroplaning mechanisms on studded autonomous nature minimizes many administrative prob- tire damaged surfaces with carefully controlled water depth lems in conducting research. and carefully measured tire and surface characteristics, and Measurement of the pattern of studded tire wear should (b) to define asperity size orders that produce wet traction be conducted in two phases. The first phase should cover and speed-dependent traction effects. relatively short stretches, on the order of 5 miles, with Road repair and maintenance hazard—Realistic analy- intensive measurements, approximately 50 per mile. The ses to estimate road repair activities resulting from studded intensive measurements would be used to establish the sta- tire wear and to project expected accident experience. tistical properties of wear patterns for barely damaged and Splash and spray—Experimental research to deter- badly worn sections. Once the statistical properties are mine splash and spray intensity for water depths between known, the sample size appropriate for measuring the en- 0.1 and 0.5 in. tire roadway can be calculated. A reasonable estimate of Vehicle lateral placement shifting—(a) In-service traf- the needed number of observations is on the order of one fic monitoring to determine vehicle path shifting over char- per mile per lane. The measurement apparatus would be acteristic intervals, and (b) experimental research to de- similar to Keyser's profllometer (166). Skid number mea- termine lane-holding characteristics as influenced by inter- surements would be made at the same points to provide mediate traffic lanes. additional supporting data. Traffic flow characteristics should be examined to estab- gathered from toll records, and to measured studded tire lish a correlation between causation mechanisms and wear damage. Particular attention would be given to the inci- patterns. The recommended study would use cameras dence of wet-weather accidents on the badly worn and the mounted on overpasses to record vehicle paths over some unworn sections of pavement. Based on recent accident distance upstream, approximately 500 feet, from the loca- experience on toll roads, it is estimated that approximately tions. The photographs produced would be interpreted and 400 miles would be needed to obtain statistically significant the results reduced to digital form. The factors to be ana- results. lyzed would be worn and unworn pavement marking, dam- If statistically significant relationships are found between aged and undamaged surface, and wet and dry weather. stud damage and accident patterns, a more detailed study of Observing approximately 500 vehicles under each one of accidents should be conducted. The study would use sup- these conditions should yield statistically valid results. Sig- plementary accident report forms to be completed by in- nificantly greater variability in the vehicle paths for the vestigating police officers. The supplementary report form damaged pavement and worn pavement markings would would detail for wet-weather accidents the nature and ex- confirm the traffic flow portions of the general model. tent of water found on the roadway at the time of the crash, and would describe the vehicle dynamics of the involved ACCIDENT DATA ANALYSIS cars. This information would be used to determine whether Accident data analysis should be conducted using digitally the pattern of wet-weather crashes is different for worn and coded information from regular police accident reports. undamaged pavement sections. Approximately one year's Using analysis of variance and regression techniques, acci- accident experience for a major throughway system should dent experience would be related to traffic volume, as provide adequate data for this analysis. 10

APPENDIX A Conceptually, the model has two main sections: preconditions SYNTHESIS OP RESEARCH RESULTS and active factors. The preconditions section focuses onneces- sarychvngem in the environsent, e.g. • presence of studded tire This section describes the synthesis of results derived from a Induced damage, which must occur before any safety related problems literature review of the degraded safety effects which result from arlme. The active factors area focuses on traffic conditions that travel over roads that have been damaged by studded tires. The ma- contribute in an immediate sense to accident causation. Within terial is organised so as to emphasize the rear-round effect on these two general concepts, the problem or wet skid and dynamic driving safety which results from the winter use of studded tires. hydroplaning was illustrated under the label of "main effects'. Both safety and wear factors are treated; the latter only to the Since this topic and others are amply covered in the following extent necessary to provide background information for the discus- sections, the disoummion here is confined to general cosssents sions of safety effects. The emphasised safety-related effects are which were useful in synthusicing the various areas. those which result from pavement and marking wear. These are listed The predominant precondition is exist ence of studded tire as below In the order of decreasing importance: induced pavement damage. Test data and field observations amply Tire hydroplaning and net skid. demonstrate that studded tires will damage road surfaces (5, 29, Pavement maintenanr.e hazard. 95, 164, 166, 153). A secondary precondition is obliteration of Splasn and apray. pavement markings, which seems reasonably well established although A. Vehicle tateral psoceventshift1n. onsamaller research base (99, 103, 181). The main problem then Adverse transverse forces and steering inputs. is to indicate when the preconditions are likely to occur. Driver fatigue resulting from noise and vibration. Pavement wear is a function of several ractors which include Ejected studs thrown from high-speed vehictes. the volume of studded tire equipped vehicled, the volume of B. Velsic la cav:ponen t iera dalton. other traffic, construction materials used, design strength of the These ranking's were based, man assessment of the relative Pavement, and climate. Normal practice is to design a pavement hazard associated with each poscible factor. 'the general yodel which will carry without fatlure an anticipated traffic volume presented in Chapter I was used to lntep,ra to the elements for some period of time. High-volume, high-speed roads are con- tua lay and to fonRvu bj vet lye maceleven ts of the relative severity structed with higher design standards and with better quality of particular proh lcu aro:aa . S Inc e It providea a use rum overview, materials and hence are lems subject to early failure through the so del et 11 be described in all nit lv more detail. rroat/water action, settling, and spmlling. Pavement wear

A-s 12

resulting from studded tire use depends on paving materials, con- specific effects in the following sections. In general, the

struction techniques, trvfftceoveventi (turning, acceleration, concept of active factors was used to describe sets of circus- normal driving), and to some extent, on vehicle speed. For etances under which particular effects might occur. The relative

particular type of pavement, the extent of wear varies logarith- ranking of effects was then etrongly conditioned on the evaluation mically with the volume or stud equipped traffic. As wear proof the likelihood of particular circumstances prevailing. Thus gresses • softer materials are removed from the pavement, and by way of example, wet mkid/hydropl aning was rated more critical the increasingly hard pavement somewhat slows the rate of wear. than splash/spray. While both involve active factors of wet

New pavements before they are cospletely cured are htghly sus- weather and high speed, wet sicid/hydroplaning presents a hazard ceptible to damage. Translated into the time domain, the under a wider range of driving maneuvers, can happen with any

characteristics of stud°imduced wear are likely to appear more traffic volume, and requires more skill to correct than the reduced rapidly and to reach a greater degree or severity on high-volume, vIsibility coming froe aplash and spray.

high-speed roads, like interstate rol4tes and urban expressways. Pavement wear and marking effacement have a ereater influence Later on, much damage will appear onelower volume roads like on some safety-related effects then on others. In addition, come

urban arterials and rural primary highways. The lowest volume effects have been studied more than others, and consequently more roads, residential streets and country secondary roads, are is known about these. The detail with which each of the effects likely to fail for other reasons before suffioiently great ia treate d will rerler t both factory. studded tire traffic passes over them. Ironi cally then, the A. 1 TIRE 5)VOROPLASIINO AND WET SKID

bett er roads are likely to be the most damaged due to their Ilydrodyninic effects are ronstdered to come into existence as

higher traffic volume. For the analysis, the rapid development vehicle soeeds on wet naveme,; ta begin to exceed about 30 moh . The

of stud-damage on the higher volume, higher speed roads placed phenomenon of hydrodynamic lift Is termed hydroplaniny'. A degree more weight on potential effects which occur more frequently on of hydroelaning is orobably involved in most wet skidding. Full these claases of roads. or dynamic hydroplaning, In whiz), the tire becomem totally seoar-

Examining the active factors presentaamuch more compli- ated from the pavement surface and nones in the manner of a water

cated picture. Active factors involve all of the rapidly mki , reeresents the list t of e lox s of con tat a friction, brought changing elements present in traffic conditions. These include about by incre,sod sneed and water deoth.

traffic volume, travel mpeeda, driver responses, and weather A three-sone conc eot conmonmy used to demo nbc the onset or

conditions. The interactions among these elements are suffi- hydronlaning is shown in Pigure A-s. The part of the tire footprint

ciently complex that they are taut described an the context of that is ass or is ted ci h nno A on the rirure Is hyci ropl snsng,, since A-3 A-4

an unbroken water film separates that part of the footprint from the pavement. In zone B, a tranaition from wet to dry traction is occurring with the water associated with zone A being forced out to

the sides of the footprint. In zone C, the traction is essentially dry, except perhaps for water globulss or other traces of water

which are trapoed in tire tread slots, or pavement interstices (2, 16. 17, 26). obviouely, before total hydroolening occurs, the con-

tact area C, and hence, the available friction force, must be greatly reduced over what is available under dry conditions.

Clearly, full hydroplaning and wet skid involve different levels

of the same phenomenon. Because the hydroplaning phenomenon is not often identlf ed as such in existing literature covering cir-

cumstances where oosplete lift-off does not occur, a coerficient s y 5 ! d :ITEX of friction of 0.1 has been selected arbitrarily In this report to

demarcate dynamic hydroplaning and wet skid. a-untroken Water Film A.l.l HYDROPLANINO. Tire hydroplaning is a manifestation of B'Transitior. State two, or possibly three, separate phenomena. oenerally, hydroplaning CDry or Substantially Dry Area is separated into that caused by hydrodynamic pressure and that due Figure A-I. :he three Lubrication Zones of the Contact Patch of a Tire Rolling or to the thin film, or squeeze film effect (1). Dynamic hydroplaning SlIding on a Wet Surface [171 is associated with water depths of at least 0.008 inches thick.

Thin film hydroplanlnp', on the other hand, Is confined to film

thlcknessea of between 3 x 10 and 0.002 Inches. In between these two regions is a range of thickness of 0.002 and 0.008 inches that

has produced what ha, been called lssinar-hydroplaning (80). La- minar hydroplaning very probably reoreaents some form of transition

between thin film hydroplaning where there is no flow in the water film, and dynamic hydroplaning where the flow in the water filmis A-S A6 turbulent.

with precsure, so as to raintaln a constant tire deflection. A.1.1.1 oynsmic Hysroplanin. As the tire encounters the Further, it was roper ted that the lydroplsntr.g speed decreased as flooded pavement in dynamic hydroplaning, a stagnation pressureue- function of tire presaurt- if the load was held constant. These re- velops at the tire/water interface. Pressure builds up in the water as the aquere of the tire velocity and, at a speed called the hydra- sul ts were modified later [1] to show that the constant deflection case had i form similar to Equation (A-i). Conclusions for the con- planing speed, equals the averago t ire-ground hsartng pressure. At stant load came remained the ssae, however. Subsequent research by thin velocity, the tire lifts off the road curface and planes acrotr Stsughton and Williams [70] indicates that Allbert' a conclusions the water. Tract tee forces due to fluid 'Ira; result inn drs are nroba biy sueDe e 5, howea.s c- , hoouu!,e or the test in:' me to CU, coefficient of shout 3.05 for water dontho of 0,1 to 0.3 inches (22). employed. Atibert 'a ha,'mnnlani rut ,.nnerlrc-ritn worn rlrr it-j'. nut Ama resUlt, the tire tasatenclr, ncy to 'cr in 'loan' and stoo rotating with the tee t tire on the out Cr our race n fir-Ct:, t icr ch-uni as Several researchers have made attempts to predict the hydropla- o onosed to thc flat trsr.k used by I!nrne . Fefi o' lion distortion ning speed by means of an analytical expression. home and Joyner (43) and the roast tveiv seell osoLjnl of alter required for lift- [23] found experimentally that for smooth surfaces and smooth, or closed patterned treaded tires, dynamic hydroplaning is very closely off evidently colored the resu]ts.

approximated by the relsttonshlp: Even after having overcome this challsngo, however, Home 'a work is stil lnot universally accepted--and with good reason. Fac- (A-i) Vd 10.3v5 tors which are obviously Important In hydroplaning (e.g., tire type water depth, wheel load, etc.) are not taken into account. In where addition, road surface and tire tread ehsracteristics are not ac- hydroplaning lift-off speed (mph). (This speed is asso- counted for. ciated with incipient 'spin-down' from full rotational Some of these factors appear in a more complete equation for speed.) predicting hydroplaning which was developed by Moore [87]. Moore's p tire Inflation pressure (psi) equation was developed through theoretical consideration or the upward thrust provided by the change in momentue of the water layer The expression also holds for ribbed tires where the fluid depth is in front of the tire. The expression is given as follows: greater than the rib depth. Home's work was done with a large vsriety of aircraft tires, and the constant in Equation (A-l) re- (A-2) flects the characteristics of these tires. Ku,mner [27] determined where the constant to be about 13.2 for passenger car tires. The differ- sq = wheel load (ia) unce is evidently due to the dtfferent carcass mtiffnesses of the two gravitational constant tire types. p ' fluid denstty For a time it appeared that the form of Equation (A-l) was vs- B = tire tread width (in.) lid only for certain kinds of load-pressure relationships. In 1965, L 2 Allbert and walker [2] reported that the hydroplaning speed was in- LIZ 2(d)2 + (A-]) (g) - - /1 - (pg)77 dependent of inflation pressure if the tire load was Increased along and where the other terms are defined on Figure A-?. A7 A-a

(Note that the spin-up speed differs from the lift-off speed sInce, If it is assumed that as will be explained later, the former is lower than the latter. W Equations (A-i) and (5-5) are, therefore, only indirectly compera- p = SwX (5=11) ble.) The basic difference between Equations (5-2) and (5-5) is that the various factors associated with wheel dimension, contact then Equation (5-2) reduces to Equation (A-i). Thus, Moore's equa- length, etc., are lumped into a single lift ooeffioient term. tion be more comprehensive in including the various factors which Tread effects are also included In this term. contribute to dynamic hydroplaning. It is not accurate, however, in In addition, surface effests are also implicitly inoluded in that hydroplaning speeds predicted by the equation are about twice 01 although the pro portion contributed by the surface is unknown, as high as those occurring in practioe. The discrepency is due to since all tests were made on the same surface. Water depth was effects in the contact area--specifically, the sinkage cone (lone A measured from the top of the surface asperities, however, which to of Figure 5-1). some degree remove d surface dependence. Under wet conditions the tire floats on a water film in zone A. In any case, the Oengenbach equation seems to be the best yet The thickness of the water film decreases progressively toward the available for predicting incipient hydroplaning. This equation will rear of the contact patch. If the time of sinkage of a particular be used as the main hydroplaning prediction equation In other sec- tread eiement exceeds the traversal time of the element through the tions of the report. contact area, then zones S and C do not exist, and hydroplaning oc- Under ordinary conditions of surface drainage, surface tex- curs. In practice, it has been found that when the hydroplaning speed is approached from lower velocities (e.g., automobiles or air- ture, tire tread, and hiahway soeed, dynamic hydroplaning is prob- craft taking off), the traversal time approaches the sinkage time ably rare (17). The changes that tire studs nroduce in surface when the hydrodynamic upward thrust is but a small fraction of wheel texture, thererore, may not strongly influence dynamic hydroplaning. load. Thus, it appears that an equation like (5-2), but modified The channels or trouchc that are cut by studs, can, on the other for sinkage effects, is required. The main stumbling block Is that hand, be cuscecced of trspnim:'suffacier,t amounts and depths of sinkage time is also greatly influenced by surface roughness and u.'atercolrcreeae the probability of dynamic hydroplaning. tire tread pattern. All.? Thin Film dydroplaning. Thin film, squeeze file, or A partial solution to the problem, at least in terms of tire viscous hydroplaning is the result of a smooth tire being unable to effects, has been provided by Gengenbach [11, 12]. Gengenbach has penetrate a thin (less than 0.002 inches of water) but tenacious produced a considerable amount of experimental data which shows that fluid film that coats a smooth pavement surface (22]. This type of wheel spin-up speed correlates closely with the equation: hyoronlsnine- differs tomriotety from that associated with hy- drodynamlo pressure. Looses in tire-to-pavement frict ion coe ffi- cient associated with dynamic hydroplaning occur somew hat more V-u 7.61,/i' (5-5) abruptly than that a500slated with thin film hydroplaning. This where phenomenon is conmidereo in more detaib by Home 1201. In general, Vu = hydroplaning spin-up speed (mph). (This speed is associa- thin film hydroplaning is a problem only on very smooth surfaces ted with incipient "spin-up' from a full stop.) [22]. = tire lift coerricient. 00th thin film and dynanbc hydrorlaning are functi onsof several factors. These include velocity, tread design, tire material, 5-50 A-SO

tire aspect ratio, tire constructIon, i,,flatinn pressure, temperature, surçace texture, wheel load, water depth, and possibly others 1351. Some of the more importent of these factors will be discussed in the following subsections. -

5.1.2 SUEFAGE EFFECTS. Pavement surface properties effect hydroplaning in terms of water accumulation. According to Galloway, at al. [2141], pavement water depth above the surface texture can be described by the following empirical equation:

= 13.38 a lO_](l/T)- (L),113(fl.59(l/5),142] - T (5-6)

where P- surface texture depth (in.) La drainage path length (ft.) I - rainfall intensity (in/hf.) S a pavement cross-s lope (ft/ft.)

From this equation it is clear that two of the factors affecting water depth are under the control of the pavement designer. These are pavement texture and cross-slope. Increasing either the pavement texture depth or the cross-slope results an a decrease in water depth. The effect of surface proper ties on hydroolaning and aet skid is best iliustrated by t:ie surface classification concept first proposed by Aiibert and Walter [2], and later modified by Ku,mner and Meyer [28]. In this schexe, pavements can be classified according to the two different asperIty s ice scales which affect friction and drainage properties: Figure 5-2. The Hydrodynamic Upward Thrust Smooth surfaces (bleeding asphalt, highly polished stone asphalt or cement concrete surfaces). Ftne-textured, rounded surfaces (worn stone or silica sand surfaces of fino gradation). Pine-textured, gritty surfaces (new silica sand or metal carbide-epoxy surfaces).

5-9 A-t2 13

Coarse-textured, rounded surfaces (polished slag or limestone surfaces or large gradation or uncrushed gravel surfaces). Coarse- textured, gritty surfaces (new slag pavements con- a!:; t ing of large part. Ic! os p ossesstn g large- and small- scale macroscopic roughness, or limestone surfaces whidh contain more than 10 percent sand-sized siliceous material). The gritty or harsh microtexture of a surface determines its basic ------friction properties and hence, its low speed skid number. The ------coarse, or open macrotsmcure of a surface determines its drainage characteristics. Drainage characteristics, in turn, detersine the variation in aktd number with speed. a.5, The microtexture scale affecting low-speed aRid number is on Speed

the order of 0.001 to 0.004 in. (3, 32, 51, ii,). The canner in Figure 5-3. Effect of Micro texture on Friction which surface microtexture affects the friction level is illustrated in Figure A-3. In this figure, increased friction levers are obtained by adding microtexture toasurface of constant eacroteature. The macrotsxture asperity scale affect ing surface drainage is on the order of 0.002 to 0.10 in. [15, 60, 731. The effect of ma- i crotexture on a surface of constant microtexture is shown on Figure 5, 5-44, In the examp le shown, the friction level at low speeds remains relatively constant with increasing macrotemture. This effect is a consequence of the fact that water inertial effects vary with the

square of velocity. At low speeds, the effect is minimal, but at t co higher speeds, the result is a significant loss in friction. In- Speed -' creasing the macrotexture Increases the number of water outflow channels, and reduces the buildup of hydrodynamic pressure. Figure 5-14. Effect Of Ilmcrotexture on Friction As might be expected, these emimple 5 represent a relatively simple explanation of the effect of pavement temture on friction properties. (See meferences 47, 63, 161, for emasple, for a more thorough discussion of pavement friction characteristics.) The examples do, however, give an indication of basic trends. Whether it be the thin film or dynamic variety, hydroplaning occurs most readily with smooth tires in combination with a smooth,

A-i) A-ti

fine textured surface. For surfaces categorized as coarse or open contact len;'!!; a! ;d do fl,,'; ion,., .ea ml or i he:; r arm not independent textured, dynamic hydroplaning does not occur until the depth of of one another, however. For exss. Er, t hr wheel load divided by the the water film exceeds the height of the surface projection. For contact keen Is nominal 11' equal to nbc inflation pressure • whereas example, the hydroplaning lift-iff speed is greater for an open the contact area is a !'uie t 4,50 of both the contact length and the Is greater for tire deflect sun. Th' !;!toractio,; if reason oh ly well understood, textured surface than for a closed surface, i.e., Vd gravel seal coat than for concrete (35]. however, as cvi dence; I Ely the ouccea F Home (23] and oengenbach (11) have had in describIng dynamic hydroplaning onset analytically Measures that have been used to improve pavement surfaces re- (see Equations (A-I) and (s-S) respectively). As discussed earlier, lative to reducing hydroplaning include grooving the pavement, for smooth surfaces, smooth tires, and water depths greater than roughening the pavement ao increase its macrestructure, and adding 0.1 in., Horns found that. the hydroplaning speed was solely a func- an open porous wearing course to the existing pavement surface. tion of inflation pressure. For less restrictive conditions, Home and Joyner [23] report that grooving an airport runway with however, oengenhach desonstrmted that wheel load, tread width, tread 3/8 inch by 3/i inch grooves at two inch centers raised the nini- pattern, carcass construction, and water depth were also factors. eus water depth level, causing hydroplaning from 0.17 inches to Since dynamic hydroplaning is a manifestation of hydrodynamic lift, 0.140 inches. Additional studies of the benefit of grooving were wheel load is obviously a factor In determining Incipient lift-off, tarried out in England by Harris on wet runways at soeeds of to since it is the wheel load that counteracts the lift force. For mph (17). For a water deoth of between 0.01 and 0.02 inches, it relatively flexible t Irea • however • the tire will deflect under was found that the brslcine force coefficient was about 0.2 on an load and cause the contact patch area to increase. This results ungrooved surface, and that this remained about the same in a In the contact pressure being essentially equal to the inflation direction carallet to the rrooves. FernenL4icelir to the grooves, pressure. In the more general case, however, Oengenbach has shown the coefficient was ceamured to be ii. At a loieer sneef of 3.3 sri:, that wheel load effects are related to the square root of the hy- no di f ference was noted in braking force reel' f icienta , rsrai tea droplaning speed, and that load interacts with tire constructions and peroencliculsr to the grooves. and tread patterns to define a chmracterietic tire lift coefficient. The use of m porous wearing course to illeviate dynamic hydro- romparing smooth tires, Gengenbach found that the lift coeffi- planing Is bmaed on the principle of providing better fluid cient (OH in Equation 5-5) was :, constant of 59 for belted radial drainage in the tire-ground contact patch [21]. The porous surface tires, and that for conventional tires, C. varied according to: allows water trapped in the footprint to flow right through the 1.342 pavement as well as through drainage channels provided by the tire w tread and the macroetructure of the pavement. The wear, freeze- - 146 (if) (A-7) thaw, and microtexture characteristics ofsuch a surface would have to be considered before electing to use such a surface as a hydro- where We Is the preacr lb.' d maximum tire load capacity. planing suppression method, however. nen"nnssac', also ,Ieos'na Irs te;! t's;st c. , is gs'es tly influence:: by A.1.3 TIRE EFFECTS. The tire factors that lnfluenoe dynamic tire tread pattern. For the six patterns shown in Figure A-5, pat- hydroplaning include inflation pressure • tire diameter, tread patterns A, a, C, and 0 were found to have the same C value of 15.5. tern, tread width, tire material, carcass construction, wheel load. Pattern E was determined to have a value of 10, and F m value of

A-si A_li 14

16. The patterns were careful v cut see a:, to insure that the pro- port ion of groove voids to tho cit lie tread surface remained constant . C differences w,'rc, ti:,'::. fore, solon in, to pattern dif-

ferencer . The character sri c it, rat terns A-fl which results in a uniformly low coefficient is the open nature. of the sides of the treads. This all ows water trapped in the contact patch to be squeesed out laterally. The difference between treads U and F is attributed to the fact that water is squeeced to the center of pattern E wheress it is squeezed to the sides of pattern D. No lateral movement within the contact patch is permissible with pattern F, thus accounting for the much larger value of Cx for that pattern. It is clear, therefore, that the beet tread patterns for hydroplaning suppression are those with open passages to the tread edges which allow the escape of trapped water. A different kind of tread modification has also been used as a means of raising the speed of dynamic hydroplaning onset. This method involves cutting small slots or sipes In the tire surface In the lateral direction. Siping is a controversial matter, however, and while some researchers have reported sarked improvements in traction characteristics through siping (16, 261, others have reported siping to be of little or no value 153, 82]. Sipea are believed by some to provide increased traction through two effects acting In concert. First, the sharp edge of the alpe tends to produce a very intense local bearing pressure that is large enough to puncture the viscous fluid film aepermting the tread from the surface. Second, the sipe itself furnishes a low pressure cavity in the tire tread into which some of the now Figure 4-5. View of Staple Tread Patterns Investigated. Pyre 5.63 15. Tread depth 7mm Groove width 3scsn punctured fluid film can drain. These combined effects purportedly [ii] allow many adhesion pctnts on the tread rib to develop and thus increase traction. Home and Joyner (23] suggest that siping is most beneficial on smooth surfaces at high speeds, however, and that sipes on the sides of the tread are most beneficial. Seemingly contradictory results obtained in testing siped tires may therefore be the result of test conditions. Increasing the tire aspect ratio (width/diameter) increases

Alt the contact ,such area and causer hydroplaning at lower speeds 1351. and drag, and fluid viscosity This is due to the fact chat on aider I. ills • Zlo, water trapped In As mentioned earlier, the primary manifestation of dynamic the contact patch has alom rer escape dit.an,o and thus tnr reases hydroplaning involves the tire lifting off from the road surface the teridemcy for iiydropilil'lFiJ', (Ii. I. Tier beet colel act patch ellipse and planing along on a toid film of water. Since the tire-to- for mtntmizan, E,ydropI ant ei, therof,,,' • is one' with a maximum ground frictton is virtuslly reduced tocerounder these circum- length and amtnssium width [1]. Th.la type of eiiipse mintmicea the stances, the tire tends tospin down, or stop rotating. Spin-down water escape distance and tends to ease mice the lengths of conee has been noted at water depths as low as 0.2 mm (11]. Increasing A and 5, as dlscusled earlier in conjunction with Figure A-i. The the water depth causes a progressively lower spin-down until the importance of tire width in dynamic hydroplaning is clearly eviden- depth reaches about 9 or lc mm (35]. The fluid drag coefficient ced in Equations (4-2) and (A-5) where it is shown that Increasing associated with dynamic hydroplaning is believed to be on the order the tread width causes the lift off speed to be reduced. of 0.05 and is the result of pressure drag and fluid viscosity. Friction on wet surfaces is mostly associated with the removal The pressure drag is due to the tire displacing fluid as it planes of water at high speed, rather than with tire material. Rubber along [22]. A companion of the various fact ors contributing to the tire/road skid number in the presence of a surface water layer material it therefore, considered to be of somewhat less importance than tread design for accompiini,lnr this purpose. Braking force is given by Kumsiser [27]. cooffipienta up to 0.5 have been achieved on fully flooded, aeooth Pressure in the tire contact patch has been found to be signi- surfaces, however, through careful tread and material improvements ficantly lower in the tread grooves than on the surface of the [26]. tread ribs for grooved tires [23]. This pressure difference be- While some of the other factors are more important, tire con- comes progressively smai icr until at a speed of about Oil knots no struction can also have a significant effect on dynamic hydropla- difference is evident. ning performance (e.g., the difference in ta values for smooth, As noted earlier, when a tire has stopped rotating as the re- radial and belted tires sentioned earlier). Jr order to perform sult of dynamic hydroplaning, it use been found that a forward effectively in awet surface environaenc, the tire contact patch velocity of the tire must be reduced substantially before spin-up must not be unduly distorted as it transmits loads to the vehicie. occurs again. This phenomenon can be explained by the fact that Radial ply tires have the least distortion of all types due to the rotating tire allows incoming water to readily pass through their high modulus of lateral ri;çidity. braking force coefficient and drain from the rear of the contact patch. When the tire stops values for radial tires as opposed to cross ply tires show values rotating, however, water in-flow and out-flow from the contact 10 to 30% higher in tests on smooth wet asphalt [261. Tires with patch is greatly diminished. The water trapped in the sliding rounded (as opposed to square) profiles are also more pressure- rontact patch tends to circulate in left- and right-hand vortices. dependent relative to dynamic hydroplaning £70]. In general, the The effective fluid drag in the contact patch at this time is, lift-off speed for a smooth rounded tire can be Sc to 100% higher therefore, greatly reduced. The tire forward speed must be lowered than for a smooth square profile tire. substantially until the trapped water can effectively drain from A.1.4 auTo EFFSCTi . The primary fluid effects which influ- the contact patch and allow the tire to regain contact with the ence dynamic hydroplaning include water depth, hydrodynamic lift road surface. Trapped water is therefore the primary factor being

A'li A-2e

responsible for the difference between the spin-down and subsequent tnsrease may be enough to brIng the skid resistance up to the on- spin-up speeds (21]. ginsi broom finish friction level. The general character of the wear as it affects mtcrotexture, macrotexture and skid number (SN) Viscous fluids, i.e. • fluids that are inherently slippery, in- is illustrated on Figure A-S. duce hydroplaning more readily than ones of lower viscosity [23]. Dust and oil combinlnr with rain water produceaflutd which is more Several factors are illustrated on this figure which are of viscous than water alone and, therefore, dynamic hydroplaning interest. As the broom finish is worn away, the microtexture be- speeds associated with these conditions are iower [21,23]. In this comes smoother. This trend cont inues unti] the studs begin to same vein, stopping distances are 1.6, 2.6, and 3.0 times greater penetrate the more dense material below the broom effect and the for wet, slushy, and flooded surfaces, respectively, when compared surface becomes more gritty again. to dry surfaces [22]. The sacrotesture of a broom finish is somewhat closed, and A.1.5 ST000ED TIRE FACTORS. The eajor studded tire-induced inItial studded tire wear tends to make it more closed. As the factors contributing to hydroplaning and wet skid are pavement tex- matrix is worn away arour.d the aggregate, however (assuming in ture chinges and wear rut-C. this tile that the matrix material is sorter than the aggregate), ttesurrace becomes more open, and the pavement drainage charac- The effect of water iccumulating in worn wheel path ruts is teristics are improved, ultimately, as some of the aggregate is discussed in Section A.1.1.1. It is shown there that ruts which the macrotexture becomes slightly less open. allow water accumulations to only 0.1 inch can be the most signifi- dIslodged0 cant surface factor contributing to hydroplaning. Unevenly worn The net effect of tease two separate effects is illustrated ruts may be an additionally hazardous factor. Accumulated water in by the variation in skid number. It should be noted that this uneven ruts can produce situations there some wheeis are hydropla- skid number profile is an illustrative example for an assumed ning while others are not. The result is a degradation in the ve- measurement speed. As mentioned earlier, microtexture has the most hicle directional stability, at well as in braking and cornering effect on low speed frist ion characteristios, whereas macrotexture efficiency. is seat important at hign speed. Therefore, the variation in skid number with tie number or studded tire passes will be dependent As discussed earlier, a gritty microtexture cornbined with an upon the speed at which te skid number is measured. The profile open macrotexture gives the bea tskid resistance. Studded tire shown on Figure A-S could, therefore, be markedly different for wear has been found to be both beneficial and detrimental in main- differentmea-uresentspeeda. twining good surface skid resistance. In-service measurements of skid number variation as the re- New nertlsnd cement tenorete surfaces are cosson lv given a sult of studded tire wear have shown a wide variation in results, rough 'broom" finish to enhance skid resistance. Several resear- PyyppS [189] found a large seasonal variation in skid resistance oherm [103, 137, 1110, 180, 1831 have reported that this finish is when asphal troads were -ubected to vehicle traffic equipped with quickly worn away during initial exposure to studded tires. The studded tires. In winter, atudded tires were found to coarsen surface then becomes somewhat less skid resistant. For certain the surface and raise the coefficient of friction. The coarsening kinds of matrix/aggregate.coebinations, however, additional exposure effect was found to be most prevalent in traffic lanes which were to studded tires will cause an increase in skid resistance. The heavily traveled by heavy vehicles.

A-21 A22

For multiple lane hIghways, the friction coefficient was found to increase in the right, or slow lane as the result of winter studded tire wear. The passing lane, in contrast, showed loss in friction potential. This result was attributed to the fact that heavier vehic lea travel almost exclusively in the right lane and thus when equipped with studded tires roughen the surface more. vehicles using the passing lane are mainly passenger care and these, n being lighter, have a greater tendency to polish the surface. The 0 0 winter coarsening effect was not found to be universal, however. In b 55 some cases, both the right and passing lanes were found to have lost friction potential during the winter. In general, the changes were Number of Studded Tire Fasses not large. Otherstudies of skid resistance tend to confirm idyyppi''s results. Paveaent friction tests made in Ontario during the years asC a 1967 through 197 are sucmsarized on Table A-i [178]. The changes 0 indicated are relatively small and suggest that studded tire wear has neither worsened, nor improved the skid resistance of most Ontario pavements. Although not stated, these measurements were evidently all aide at the same time of the year. Froim,, and Corkill [159] and Smith and Schonfeld [179] in Ontario, as well as Wehner In West Oereany [183], show the same seasonal variation in skid resistance that was found by HyyppS. - Number of Studded Tire Passes A.1.6 Stil-INARY. Insumeary, studded tires have been found to be both beneficial and detrimental in relstion to cavement friction characteristics influencing wet skid. bihere a specially preosred friction surface is Ibnaded, the effect is detrisentai. On the other hand, the friction properties of a previously smooth surface SN can be imoroved significantly through exposure to studded tires. At least one researcher (i.e., Ludema [i'l]) for example, has recee- mended that surfaces of low friction can be rehabilitated by taking a hard object and soalling the pavement surface to isorove its texture characteristics. This aopesrs to be a function siready being undertaken by the tire stud. Number of Ittudded Tire Fasses While studded tire use may not have a significant adverse Figure A-S. Pavement iturfice Changes as a Function of Studded Tire Wear. A-23 A24 16

effect on pavement surface characteristics relative to dynamic hydroplanl.ng, the mace is not true of uneven wear effects which have resaited in wheel track troughing, or rutting. Ample evidence of such trmughing has been documented (164, 166, 168, 178,

o 183, 1891; therefore, the wear mechanism oroducing such effects I, will not be discussed here, the effect on hydroolaning of water a C 0 accumulating inrut depression is defined quantitatively in Sec- 0 tion All. C A.2 SAFETY ASPECTS OP PAVEMENT MAINTENANCE ACTIVITIES

Road surface repair activities are projected to be substantial Em e during the next several years if studded tire use continues unaba- mm ii ted. Past experience has ehown that the accident rate which can be b-. - EV 50 ES 0,e expected from such maintenance and repair activities can be sub- em 'a 5 'cm 5 stantial. Road surface repair activities are, therefore, considered mel '-a o Os, to be the second most important source of accidents resulting from studded tire wear. As the following discussions will show, however, 00 C SC is this feeling is more intuitive than factual, since little applies- CO C-p bte dat acre available. Such crojectiona nrovide a useful base for indicating the mm•-"-u;u 0 - eacnitade of the safety problem likely to arise from Increased mE hOc ,14a Id C 0.us repair activities. Stimi, it should be renembered that such he project ions were based on a limited number of site, many of we s,is which were rather susceptible to stud-induced damage. While the CS) 'e7 msw accun.elating evidence indicated otherwise, there is some p05- a-em sibility that the amount of wcrk needed may be mealier than as, as - currently believed'. Secondly, the amount of work required will C '0 cc denend on the degree of hazard presented by the other mafety -C an effects considered in this study. For examole , if serious hydcoplaning nrobleres aonear with 1/4 Inch of wear, far more P 5 renatr activity will be needed than if one inch reocesented a 'Difficulties encountered during the project in finding extensive sections of heavily damaged pavement in order to con duct accident F' studies caused this concern. The snotential seriousness of the :6... dl ii Bl 10 oroblem is not doubted. Rather, given the wide variatioca in deported wear, it is believed that more precision in the estimates would be desirable.

4-25 A. 26 serious hazard. the proposer] research nror'ram covered in Annen- Construction site crews contribute to accidents through ix C will allow much sore precise answers on the amount of Improper sign or warning signal placement. wear and of consequent repair requtreme nba Caut bonny remarks Sign meamagem which are incocrect and which are soon ig- notwi th Stand inn the amount of rena ir work required and the con- nored. ceqient severe hazard posed by it generates a high ranking for improper lane msrking alteration. the effect. A tabulation of construction site accidents in Illinois by aveeat surface repsirs in the State of Wisconsin 1831 are Mccarry 185, 863 indicated that the greatest hazard occurs where projected to sverage $5.1 million over the next twenty-five years the highway is under construction, but still open to traffic. as the result of studded tire damage- Approxiestely 2,800 miles of Exaxplem included bridge replacement or pavement connection pro- pavement are involved- In Minnesota, the costs are expected to be ject a where it was necessary to build a detour around the original $2.8 million for 1973 and are projected to rise to $13.3 million alignment and thum introduce sharp turns. per year by 1980 [12). Costs in, the Province of Ontario are esti- The next most dangerous situation involved patching, widening, mated to total $127 million through 1979 for repairing mtud-damaged and resur facing, where the road could remain open but where barri- pavements and pavement markings (103. 91. Prom these estimates it cades were required intermittently. Most accidents in these cir- can be coneluded that roadway maintenance activities will increase cumstances involve d rear-end collisionm. Drivers failed to recog- substantially over previous level- as the result of studded tire nise in time that traffic ahead was stopped, or slowing down. It .,or. Tb.OsconclusIon Ic. supported by projected repair activities ia this type of maintenance activity which will be prevalent rela- 'or Wisconsin (833. As the result of studded tire damage it is tive to studded tire-dsimaged road surface repair resulting from estimated that, over a period of twenty-five years, as aany as ten studded tire damage. additional pavement overlays will be required for stretches of Construction site accident statistics for the States of Texas, heavily traveled roadway. Virginia, and htyoeing are riven in Tables A-S, A-i, and A-i, rem- Several stu dies have shown that tnreamed accident levels can be esrected when maint entc.ce are construction activities are in nectively (i7]. A more detailed break down of Virgir,is atatistics pcogrsss. Sample accident mechanisms include; for the six-month period between May and October of 1968 is given Coflac Ion with maintenance equipsent. in table A-S (88]. It is clear from these tables that construc- Collision with other vrhlcles tion and maintenance sites r,enerate substantial numbers of acci- Driving lisics work ni-ca,,. Los,r of control hrrau,sr uf road arcrfacr conditions. dents. While these craahes are a relatively small proportion of in een.rnl , construction site accidents occur through error on the the total, lccrcent to inercent, they probably reeresentan part or motoriste, or construction crews. Driver err Dcc may result extremely high crash rate since in all likelihood s much srslier from; portion of vehicle miles are driven under construction conditions. ConfusIon in reading signs, or warning mignals. Not following eigc]-indicated messages (e.g., slowing down). Being confronted with an unexpected situation.

4-27 4-28 17

turther, based on our faniliarity with the Texas esnerience, con- f:ihlo A-I struction site accidents apneas to he more severe since they ren- TEXAS: Rural Motor VehI c 5:. Accidents at Areas Under resent a hirher fraction of fatal than of nonsatal crashes. These C on S t ruct Ion * observations must be somewhat nualified since data are tncosnlete Year Fc.tal Accidents Nonfatal Accidents and not collected on a national basis. 1960 60 1664 In eussnary, cost projection analyses have shown that road re- 1961 49 1598 pair activities will be significantly greater as the result of 1962 50 1450 studded tire damage. Further, construction site accident rates in- 1963 56 1517 volving resurfacing are aeong the highest for all maintenance acti- 3964 53 : 2131 vities. With the continued use of studded tires, then, it can be 1965 30 2033 expected that road repair activities will be one of the highest 1966 51 2341 sources of associated accidents. Factors which could affect this 1967 54 2650 conclusion are: improved construction site accident prevention measures tSourco: Texas Depart sent of Public Safety, Austin, Texas Improved road surface and etud design characteristics which result in reduced wear. Revised road wear projection data. Table a-I A more definitive conclusion relative to expected accident rates VIRGINIA: Accir.ents at Areas Under Construction would require the acquisition of more refined construction site acci- All Fatal Personal Property dent data and the updating of projected resurfacing activity esti- Year Accidents Accidents Injury Damage sates. 1960 574 : I4 1961 881 5 217 659 A.3 SPLASK AND SPRAY 1962 971 15 247 709 Splash and spray are terms that are applied to water that is 1963 1509 7 272 830 ejected to the side, or behind a vehicle as it travels over a wet 1964 1319 9 307 iaa pavement. Splash is usually defined to be the larger drops of water 1965 sIlt 16 383 11415 or slush which fall to the pavement iemsediately after becoming air- 1966 1763 13 3941 1356 borne. spray Is a term given to the finer droplets which become 1967 1734 9 4117 1308 entrained in the turbulent wake of the vehicle. Measurable amounts of spray can result from water depths as thin as 0.01 inches. tSource: Vircinia Praffit Crash Pacts, plus information from the Clearly, stud-produced trough swill contribute to splash and spray Virginia State Police Hetdquartera, Richmond, Virginia problems.

A29 A30

Departure snobs Ire on Inc order or Ic dogrees from the horicontal. Table A-A Spray droplets are liftea off the road by the tire and carried along WYOMING: Accidents at Areas Under constructiont in the turbulent wake of both the tires and the vehicles. Water contasned in tire groove.; and sipes is thrown clear after about one- Total Year Number of Accidents Injuries Fatalities quarter revolution. The aaount of water thrown free decreases as the tire continues to rotate and essentially stops after two-thirds 127 76 3 1962 of a revolution. Wind and vehicle speed tend to force the droplets 108 54 4 1963 backward; the greater the relative wind, the more tendency for a na. na. na. 1°6l splash drop to shatter and become part of a spray mist [96]. isiS 100 95 3 1966 118 86 6 Accidents resulting from splash and spray are relatively uncommon. An analysis in cngland indicated that 1.3 percent of wet Sou,cr: The Wyoming 11196c construction Area Rotor Vehicle Acci- weather accidents are the result of mud or water thrown on the dents" (discontinued after 1966). Among the material s surveyed this windscreen [97]. A similar study in the state of Michigan of both one presents the most elaborate analysis of accidents at construc- wet and dry weather accidents has shown that the accident frequency tion sites. It coapares districts • gives descriptions of the acci- resulting from splash and spray is on the order of 1 to 500, or dents, and concludes that 32$ of the accidents occurred entering or 1 to 600 [92]. Evidently, accident" resulting from splash and leaving detours with 13 such accidents assigned the cause. 'Failed spray occur most frequently during overtaking maneuvers or during to make curve onto detouroverturnedm. passing maneuvers from the opposite direction. It is not completely clear what the accident producing effects from splash and spray are, however. It appears that drivers tend to learn to compensate for the effects of splash and spray. When accidents do occur, the ori- Table A-S gin seems to invo lveasudden loss of visibility which causes the of Accident Data driver to carry out some type of panic maneuver. Roadways randomly construction rutted through the use of studded tires would obviously be more Construction Equipment Maintenance Total likely to produce these conditions. The factors influencing splash and spray include water depth, Total Accidents 513 74 93 680 vehicle speed, tire tread pattern, road surface characteristics, Persons Killed 17 0 5 22 and vehicle characteristics (i.e., shape, fenders, mud flaps, Persons Injured 242 18 57 317 wheel arrangement, etc.). Each of these factors will be discussed Property Damage $375,550 $53,785 $80,370 $509,695 in terms of its individual relation to splash and spray in the following subsections. A.3.1 WATcH OEFTH. Water film thickness seems to have little Splash and spray laboratory teats on an uncovered wheel [95] influence on spray but greatly influences splash [94, 951. As in- indicate that the major portion of water is displaced sideways in dicated earlier, depths of 0.01 inches have produced significant the form of splash. Water thrown off in the rearward direction splash effects, while depths on the order of 0.001 inches seem to leaves the trailing edge of the contact patch in the form of spray. A-32 A31

be the scallest that produce any significant amount of spray. The A. 3.3 ROAD SURFACE. Cones-ni. 10,51 road surface:; may or mey not functional relationship between splash production and water depth is have an effect on spray production. Surf aces made of impervious evloently unknown, however- gravel and chip dressings were found to reduce spray by as much as a In an tnvestigation on the effects of mud flaps on passenger factor of two over untreated surfaces [97]. Control in these ex- cars, Chapoux [91) round that spray length ancresses approximately periments was not sufficient to guarantee results, however, and the as the square root of water depth and as the square of speed. Spray effect of surface treatments on splash and spray production is there- height, on the other hand, evidently increases directly with both fore not clear. velocity and water depth, but does not exceed the height of the An experimental porous surface was also examined under these vehicle. same wetting conditions and this surface produced no spray whatever. A.3.2 VEHICLE SPEEJ. Vehicle speed has an effect on both A characteristic of a porous surface is that it allows water to be spray density and the production of spray through shattering of drained down through a porous layer without the necessity of drain- ing along the surface itself. While alleviating problems with splash droplets. According to Maycock [97], spray density is insignificant below 30 mph, but increases rapidly thereafter at about splash and spray (also hydroplaning) the porous surface lacks dura- the 2.8 power of speed. Most water displaced by the tires below bilIty and tends to lose porosity with time. 30 mph falls to the pavement without breaking up. According to A.).) TIRE TREAD PATTERN. For water depth less than 1/8 inch, Lane [98], splash droplets shatter into spray according to the re- Mayccck [97] found that a smooth tire produced more side splash lationship than a treaded tire while the spray differences, were insignificant. Other research has shown that side splash is progressively greater (u-v)r = 54 (A-8) with (1) winter treads (I) summer treads (3) all tires. Raise,, at ml. [95] found that sidesplash could be greatly reduced by bonding where flexible chine to the tire aidewall. In the same set of experi- u_vwmelative drop velocity in mph- ments, it was found .that a transverse tread is best for removing Droplet diameter in inches. droplets from a tire and that it 3m also belt from the hydroplaning For dropa between 0.05 and 0.10 inches in diameter, the critical suppression standpoint. In this last context it has been found that speed for shattering is between 20 and 35 mph. This effect isrea- a hydroplaning smooth tire produced the greatest amount of splash ponaible for the large increase in spray (mist with increasing [914] speed). A.3.5 VENICLE FACTORS: Maycock [97] found that vehicles with Spray patterns are influenced by wheel speed in that spray sloping tacks produced about twIce the spray as did vehicles with thrown forward increases with increasing wheel speed. Splash drop- flat square backs. Kvidently vehicle aerodynamics has a great in- lots leaving the tire tangentially in a vertimai direction move fluence on the pattern of spray while the spray density is mostly forward at vehicle velocity. Those leaving later move at a greater influenced by tire type, water depth, and vehicle speed 1951. velocity than the vehicle and, hence, have a greater tendency to A.3.6 COUNTERMEASURKS. Measures to counter the effects of shatter into Spray mist. Smaller tires traveling at a greater ro- splash and spray have gener ally been concerned with reducing or tationalapeedwould, therefore, tend to produce more spray than removing the source of the a irbos'ne water. Since most splash and would larger tires for the same vehicle speed. A-I 4

spray problems arise from coassercial vehicles, most of the applicable A.4.1 LANE MARRING EFFECTS. Research by Gordon on a two-lane research has been directed toward developing fenders and mud flaps road (one lane in either direction) [100] has shown that the road edge for splash and spray suppression [1117, 183, 95, 971. Carefully de- and center line are the primary visual in puts which the driver uses signed mud flaps, for example, have been developed which reduce spray In guiding his vehicle. These two cues involve between 80 and 90% of behind a large commercial vehicle by as much as a factor of three or the driver's eye fixation positions in the normel course of driving. four. Other spray suppression devices have included a hood mount It would seem logical to conclude, then, that intermediate lane mar- detector which is used to deflect airflow away from the windsireen, kings on either side also provide the primary vehicle guidance stimu- Spray reductions on the order of two to ten times have been experien- lus on multi-lane highwaya. Lane marking obliteration should, there- ced with such devices [97]. Problems have resulted with hood mounted fore, produce degraded lane holding and vehicle guidance characteris- deflectors, however, in that the dropa tending to flow around the tics--particularly on curves. deflector and land on the windscreen tend not to merge and produce Before and after accident from the state of Ohio indicates that mottled effect which Is difficult to see through. pavement edge markings along rural two-lane highways have made a sig- A.3.7 SUMMARY. In summary, it can be understood that splash nificant reduction in accident statistics [102]. The data indicated and spray effects are reasonably well understood and that studded a net reduction of 37% in fatalities and injuries with a significance tire-induced pavement ruts will increase both the frequency and in- at the 0.02 level. tensity of windscreen visibility loss. The increase in subsequent Other studies have shown that pavement markings have a signifi- accidents rates is problematical, however. Heretofore, accidents cant effect on vehicle lateral placement. Data collected by Willis- resulting from splash and spray have not been a significant factor ton [105] at four different mtrsight, two-lane sites show that the in accident statistics. It is expected, therefore, that splash- and lateral placement distance from the centerline becomes progressively spray-produced accidents are a less probable source of studded tire- larger as lane marking cues become fewer. The data are sumcsarized on induced accidents than the effects discussed in Sections Al and Table A-i. The data also show that placement distance is influenced A.2. by the Presence of opposing traffic. The most significant edge marking effects on placement were found to occur at night. After dark A,4 LATERAL PLACEMENT SHEPTINO CAUSED BY MARKING WEAR AND WORN WNEEL PATNS the majority of drivers tend to drive nearer the middle [104]. Vehicle lateral placement shifting can result from pavement Table A-i Lateral Placement From Centerline (ft. ) As Influenced marking obliteration and wheel path wear. The former results in a by Pavement Markings loss of delineation lines which the driver uses in guiding his vehicle in lane holding, while the latter produces several adverse MEETING FREELY MOVING OPPOSIN0 TRAFFIC effects which may cause the driver to steer to one side of the worn Site Markings 1 2 3 4 1 2 3 14 paths. Among these several effects are: (1) retained water in the path ruts which can freeze • cause hydroplaning, wet skid, or produce Centerline & Edge Lines 2.6 2.9 3.14 14.1 3.0 3.5 3.8 14.6 splash and spray problems; (2) roughened pavement which can cause Centerline Only annoying vehicle interior noise and vibration levels; and (3) ad- 2.7 3.2 3.7 3.2 3.1 3.5 4.3 3.5 verse steering effect, from the tendency of tires to climb the rut No Markings 3.2 1.0 44.3 3.2 3.8 414 44.5 3.7 side slopes. Similar studies by Hubbell and Taylor [101] tend to confirm, these conclusion,. It was found that the extremes in lane holding A-li A-36 19

were represented by no markings and a solid yellow centerline plus edge lines. Edge delineation was round to be the only significant Table A-? element on all cases. Placement Comparison Bstween Worn and Unworn Pavement Sections While it is clear that lane markings can have a significant Distance From Vehic]e.Centerline to hight Pavement Edge effect on both accident causation and lateral placement, no survey vehicles has been identified which shows the general prevalence of marking An'! t,hmeti c ltar,darcl Coeffi cient Within Three rican Devi,tion of Variation Foot Wheel Ruts wear as the result of studded tire use. LANiS WITH wJ!':IiJ. 6.01 ft. ] .17 Pt. PATH WEAR (19) - 21.23% 73% A.4.2 OTHER EFFECTS. The occisrrence of other effects which may lead to lateral placement shifting such as water accumulation TJNWORH LANES (7) 4.26 ft. 1.12 Pt. 21.16% 77% in ruts, roughened surfaces, and adverse steering forces and mo- menta are all discussed elsewhere in terms of individual contribution as accident causation mechanlsma. H o connection has yet been found, however, between these effects and lateral placement shift- accident rate.s are lane ielineotors--particularly edge markings. ing. In the only known study which has been made to determine These are not located in the areas of substantial wheel path wear lateral placement effects resulting from studded tire wear, the re- along most highway .seetiona, however. Therefore, marking wear mus ts were negative [99]. In these studies, eleven straight-away cannot he idants Pied as a cossonly pre valent sourceof lateral sections of freeway with at depths ranging between 0.26 and 0.56 placement shifting. On the other hand, lateral placement shifting inches were compared in terms of lateral placement with fifteen for nther reasonc. could .ecd to marking wear and thence to a com- similar unworn sections. Videotape units recorded vehicle place- pounds ng ffrr t or lid 11,5 nail plaeosent shifting. No evidence sup- ment distributions at specific point locations on each pavement per ti ng thin hypoth. a i - :,ic. been prod Lice'!, however. section. A synopsis of results is shown in Table A-?. The simila- rity of results. indicated no cignificant or discernible differences Marks rIg wear has been repnrtod to be uncosesoely excessive in in vehicle placement between freeway lanes with appreciable wear areas where lane croa'.;ln.'; and weaving it prevalent [103) (e.g., as compared with lanes not exposed to studded tire wear. It should curves, or interchanc'.o rump locations on sulti-lane highways). To be noted, however, that all surveys were made in the presence of compound the problem, lane mark'nc at t.ase locations are of dry pavement and in clear weather. Therefore, no conclusions can greater need in vehicle ruidonco and their loss is therefore of be drawn reistive to avoiding wheel ruts where trapped water is pre- greater significance. suco location; soul drepresent an exception sent. to the general lack of a cause-effect relat ions hip between studded tire wear and lateral p3acemer.t shifting. A.4L3 SUWRY. It is evident that lane marking obliteration can have an effect on vehicle guidance and accident causation. In Other than lane aar,-ring obliteration, no wear effect has been the absence of survey data showing actuai'marking wear patterns, shown to cause lateral Ptacementahifting. however, It would generally appear that marking obliteration is not A.5 ADVERSE TRANSVERSE FORcES AND STEERrNG INPUTS a large studded tire-produced accident causative factor. The pavement markings which are most important in affecting placement and Adver-e transverse :'orces and steering mosents may be transmitted to a vehicle as toe resalt of tire interacti on with studded A-37 A35

tire-worn wheel ruts. The adv,ce:le c'rfocts could result from two that of the ,'rrects LIlacu.';rrd , -arli"r. mcchaniema AS.? TIRE/RUT EDO-i INTERACTION. Tire;, in general, have a Roadholdlng degradation resulting from a roughened surface tendency to climb the side of a ridge or slopIng surface. In the which may occur during cornering and/or braking. tire trade this phenomenon is known as "nibbling'. A climbing Tire/rut edge interastion as the result of tire nibbling reaction does not always occur iewsediately upon contact with the characteristics, ridge, however. Initially, most tires running up against a low A.5.1 ROADHOLDIHO IFFECTS. Roadholding effects are influen- vertical ridge will tend to ehold off' until sufficient side force ced by pavement asperities which are of large scale compared to is produced to cause the tire to bite into the ridge face. Once those responsible for producing tire/road friction. Such asperi- climbing on the ridge is initiated, however, a side force tending ties are aeasured in terms of inches, whereas friction-producing to reinforce climbing quickly builds up. macro- and microtesturea are measured in terms of fractions of an Marshall at at. [lcOJ show data which illustrate these cha- inch. racteristics for three kinds of tires. The tires were made to While the friction-producing characteristics of a pavement croas • a 1/2 inch plank at 5 1/2 degree angle with respect to the surface can be described in terms of a skid number, this number may direction of travel. An initial "hold off" tendency followed by a not be reflected in the maxisue cornering and braking forces which force tending to reinforce climbing is clearly evident for both can be generated by the vehicle. Such forces are a function of bias and bias-belted tires. The saae is not true for radial tires, vehicle speed, suspension, tire characteristics, large scale road however. In this latter case, the forces are generally in a direc- surface asperities, road surface contaminants, and the particular tion to oppose climbing. The peaks and valleys in the nibbling characteristics of the maneuver. In order for a vehicle to take responses evidently correspond to ribs and grooves in the tire tread. advantage of the available pavement friction, the suspension sym- The data of Marshall et si. [109] show that the side force temmust act to keep the tires uniformly in contact with the road quickly varies from -125 lb. to +135 IN. for the bias and bias- surface. Large scale surface asperities--those important in vehi- belted tires. This 'swing back" tendency has important implica- cle considerations--tend to degrade this capability. In tions. If a driver tries to steer across a ridge, the initial res- addition, a particular cornering or braking maneuver may result in ponse of the tire is in a direction to oppose his actions. For the unequal tire loading which Is independent of large scale surface radial tire this oppoattiors tends to continue for the duration of effects (e.g., vehicle mass transfer through pitch, or roll). the nibbling procees and provides the driver with a consistent and While a good deal of research has been done in these areas, predictable feeling. For the bias and bias-belted tires, however, [106, 110, 111, 1129, mush is still not clearly understood. In the 'swing back" tendenctr tries to pull the tire across the ridge particular, no research hia been identified which shows the effects and can ccae as a surprise to the driver. Drivers trying to get back of traveling over worn ruts while at the same time carrying out onto the road after going off on the shoulder have been known to braking and cornering maneuvers. Therefore, in the absence of such spin-out because of this effect. information it is not clear what effect ruts and roughened sur faces The implication of torques being applied to the steering sys- would have on these asneuvers. Intuition would suggest, however; tem through nibbling is not considered to be as serious. A torque that the resulting accident causation hacard is not as large as applied at the tire is typically reduced by a factor of about twenty

A-3e A45

by the time it. is felt by the drIver. The resulting disturbance experimental evtdenee is available which supports the existence of either phenomenon • howevur, If pavements continue to be worn at torques of one or two foot-pounds, applied at the steering wheel, are well within the control capabilities of all drivers - Further, present rates, the liketlhood of roadholding degradation, and ad- for cars equipped with power steering, the disturbance torque will verse rut travel effects will be increased. In this event, vehicle trsveling over roughened and rutted be negligible. It can be concluded, therefore, that the tire handling characteristics it, aligning torques produced by nibbling will be of less consequence pavements should be more carefully esamined. In present circum- than the resulting side forces. stances, however, adverse transverse forces are considered to be of less consequence than the effects mentioned heretofore in terms of Other data showing nhc nibbling characteristics of tires tra- accident causation mechanisms. vel ing over objects is presented to [107. 113]. In contrast to the results for nibbling, their data chow that radial tires (rigid Rut or roughened eurfsce.-lnduced steering torques are consi- breaker) are more influenced by the obstacles than conventional dered to be of negligible consequence. designs. whatever the relstive difference, however, it is clear that tires running over road irregularities can generate signifi- A.6 NOISE AND VIg5SATI0N EFFECTS ON ORIVEN FATIGUE BESULTINO FROM BOIJOHENED PAVEMENT cant side forces. The subjective effects of noise and vibration are not eonsis- The implications of these findings for tires traveling over tent and depend upon the subject, the setting, and the type of wheel ruts produced by studded tire wear is not completely clear. disturbance. In general terms, noise is heard and influences the - cannot be es- Studded tire-worn ruts have sloped a ides and hence body through the ear, whereas vibration is felt, affecting scan pected to generate lateral forces which are as large as those pro- primarily by the wideapread mechanical disturbance of organs and duced by ridges- The forces may be large enough to draw a vehicle tissues - High intensity, low frequency sound (less than 100 hc) to one side or the other of a rut, however. If the driver.attempts can enter the body by direct absorption through the body surface, to steer no as to remain in a rut, a side-to-side jostling effect however, and can excite non-auditory sense organs. In this Connec- can be expected. The severity of the ph enomenon will be a function tion, the ultimate physiological effect of such noise is essentially of rut profile, tire type, vehicle characteristics, and the deter- sieilar to that of full body mechanical vibration. The beat known eination of the driver to remain in the wheel paths. The unpleasant frequency dependent effects of noise and vibration on man are re- effect could be ciaply alleviated, of course, by the driver shifting ported by Ouignard [122]. his vehicle to one side of the worn paths. In fact, a vehicle left to itself would tend to climb out of the worn paths on its own The separate effects or noise and vibration on driver fatigue, accord- As indicated in Section A.4, however, no evidence of lateral as produced by road roughness, are discussed in the following sub- placement shifting in practice has yet been found. sections. A.5.3 SUWIARY. Adverse transverse forces may occur as the gil NOISE EFFECTS - result of roughened paveoent sur faces, or worn wheel paths. The A.6.1.1 Sound Measurement. Sound level is eomseonly measured rorner may affect vehicle roadholding characteristics, and thus cor- In one of two ways. Sound pressure level (SPL) is an absolute mea- nering and braking. The latter may cause an adverse aide-to-side sure of sound pressure which is defined as follows: motion when a driver attempts to travel in the worn ruts. No direct

A-42 A-4 1

noise control are suggested by Bosenbitth, et al. [131]. ira = 10 log PC (A-9) ,Of Table A-B Limits for leafness-Pvoldance and Comfort In other words, SPa is the root mean square pressure, P. related in Nisimum Permissible Sound Pressure Level, decibels to as reference pressure. 5'ref'- In general, Perf is equal above 0.0002 microbar. to 0.0002 eicrobsr. The units of SPL are referred to as db(C) when oeafnesa-Avoidance Criterion Comfort Criterion read on the 0-scale of a sound level meter- The C-scale represents Repeated a response which is essentially unifors over the audible frequency occasional Eaposure Octave Exposure (Period of Noisy (People Quiet (Feople spectrum- Band (1 Hr or Less) Months) Espect Noise) Espect Quiet) georecomonly used sound pressure measurement method is 38- 75 125 115 100 80 based on the A-weighted scale equalisation scheme. This weighting 75- 150 120 110 95 70 scale approalmates the inverse frequency response characteristics 150- 300 120 110 90 60 of the human ear at the 40-phon loudness level [119]. Little weight 300- 600 120 105 85 55 is given to low frequency sounds. In general, the A-weighted scale 600-1200 115 100 75 50 is statistically indistinguishable from the beat psychologically 1200-2500 110 95 65 50 derived measures in its relationship as a predictor of human res- 2400-4800 105 90 60 50 ponse to vehicle noise. 4800-9650 110 95 55 45 Careful attention must be given to differences in the meaning between db(C), which is a direct measure of sound pressure level Annoyance, fatigue, discomfort, and irritability ire frequent and db(A), which corresponds to the A-scale weighted sound level subjective reactions to exceaa noise [1351). Mental performance and measurement. For essmple, a change of 10 db(C) In SFL means that motor processes are influenced by noise--at least temporarily £125). two different sounds vary in power by a factor of 10. On the other work carried out under noisy conditions requires a considerable in- hind, a change in 10 db(A) on the A-scale means that two sounds crelae in energy expend iture even though output may remain the same. easentially differ by a factor of 2. A difference of 20 db(A) means In oomplea tasks, performance falls off under high noise conditions that one sound Is rated as four times as noisy as another [120]. (a slight effect is noticeable for noise above the 90 db(C) level) [95]. oeoiaion tasks also show a performance decrement at about A.6.1.2 Subjective Response to Noise. In the general subjec- the 90 db(C) level. The decrement Is in terms of more error, rather tive response to noise, intensity and duration as well as frequency than as a reduction in output. This is also true for simple or re- are important. Diecosfort is not a reliable guide to loss of wor- petitive tasks. In ihort, noise can have widely deleterious effects king efficiency or impsireent to health- Nan can find pleasantly on task performance. exhilarating a level of noise which will cause permanent hearing damage [122]. peafness avoidance and comfort criteria as a func- On the other hand, some types of noise can lead to increased tion of noise frequency are shown on Table A-O [1251. These are work performance. For example, intermittent and impulsive noise suggested criteria for autcmcttve designers. Other criteria for may be beneficial to persona involved in boring tasks. For simple,

A'-43 A-44 21

low information processing tasks, noise tends to Improve performance The effect of a roughened road surface on vehicle interior [lii]. Steady low intensity noise, however, usually produces no noise has been determineo from data collected for 29 different change in performance. automobiles traveling over the SOFA stone-aett surface (a level A.6.1.3 Vehicle/Roadway Noise Characteristics. The amount of cobbled surface [126]). The mean noise levels resulting from tread and tread pattern, roadway roughness, wetneaa, stiffness of traveling over this surface are about 8 db(A) higher than compara- the tire casing, tire loading, vehicle speed, and the coupling be- tive values for the tarmac surface. The highest values are on tween tire and vehicle body are all important in determining the the order of 98 db(A), however. These values exceed the noisy amount of noise produced by a vehicle traveling along a roadway comfort criteria, as discussed in Section A.6.1.2, and are at a [119]. Accurate data on the noise produced by the various tire/road- level which will cause some fall-off in driving task performance. way interaction mechanisms for current tire/roadway combinationa are water on a pavement surface may increase the resulting noise unavailable, however [120]. In addition, although the various level produced by an automobile anywhere from 3 to 15 db(A). For components involved in vehicle noise production are recognized, the eaample, wet smooth asphalt is 15 db(A) noisier than the same specific mechanisms are not well understood, or easily measured surface when dry, while a rougher portland cement surface is 8 [136]. Noise produced by tires, in partirular, is not easily alle- db(A) noisi er when wet [129]. Evidently, the rougher the surface, viated. Factors which increase safety, traotton, and tread wear the lesser the increase In noise from dry to wet conditiona. in tires also contribute to increased noise [133]. Mills [126] indicates that road noise i C more sensitive to the Typical noise spec tra eeasu red maids two vehicles with the presence of water at low speeds than at high apeeda, and at low windows closed are reported by callow [ill]. These data were ob- vehicle weights, rather than high weights. tained during travel at 145 mph over a smooth, dry, tarmac surface. AG.!), Noise Produced by Studded Tires, and Pavements Peak noise levels are 744 db(A) for an Austin Mini Mk. II 850, and Roughened by Studded Tires. The noise produced by studs on 69 db(A) for a Ford Zephyr 6 Automatic. For these vehicles, the studded tires is, in general, insignificant [133], A slight buz- noise levels are just at the dividing line for Speech Cosanunication zing is heard when both front and rear tires are equipped with Level 55 [031]. Although these data are tynicas, they do not studs. Stud noise is almost unnoticesble, when only the rear tires necesearilv represent the extremes. are studded. In general, stud noise is heard most clearly by a Peak aound level tends to increase as a function of vehicle bystander, at low speed, on a smooth hard surface. speed [117]. The data are a cocposite of seaaurementa from 214 dif- The noise produced by surfaces roughened by studded tires ferent automobiles and attain represent travel over a dry, smooth very probably lies between the noise produced by the smooth tar- tarmac surface. mac and stone-sett surfaces described in the previous section. A Callow [117] shows that maximum sound levels are about 87 relative comparison of these two surfaces along with a apray and db(A) near 70 mph: Since the spectrum shape for vehicle noise is chip 0/14 in. chippings) is shown on Figure A-? [126], The spray relatively independent of vehicle speed [119), it can be concluded and chip coating is about midway between the other two in terms that these pea Ic values occur at frequencies somewhere between 75 and of noise-producing characteristics, and is probably comparable to 300 ha. Therefore, even at 70 mph, vehicle interior noise, resul- a studded tire-damaged surface. It would appear that vehicle in- ting from travel over smooth surfaces, is below the noisy comfort terior noise, resulting from travel over roads roughened by criteria.

A-al A-46

studded tires • would be _n the 90 db(A) range. Alt1iugh the results are highly suhjecttve , this is a noise ,evel s:sich can pro duce sap- ginai long term drtver fatigue. A.'6,2 VIBRATION EFFECTS A.6.2.1 Subjective Response to Vibration. Repeated attempts 90 have been made to define limits of human exposure to mechanical vibration. Most of the appiicat ion has been in transport or military vehicles where driver comfort or crew performance has been the criterion. The difficuitiea sasooiated with defining limits are 80 primarily due to the fact that m,asured res ulta must be based on subjective reaction. As cent ioned earlier, such reactions are not consistent and depend upon the subject, the setting, and the type a V of dis turbance, In bro ad terms, human tolerance to whole body vi- 70 iration is iowest in the b-S hz band of frequencies [122]. This is recognized in the efatigue_decreoaed proficiency limits for vertical vibration' which have been proposed by the International Stan- dardltatlon Organization (ISO) [121]. The 14-8 hz frequency band 601 I I I 1 II evidently corresponds to visceral resonance which causes these fre- 20 30 140 50 60 70 80 quencies to be highly ohectionable. In general, up to about four Road Speed. MPH. minutes exposure, time is not a factor in the ISO limits [123]. Prom that point up to 214 hours, however, the acceleration tolersnce - Dry So.coth Tarmac imita are reduced cenrlr:uouaiy to a level of about 14% of those Dry Spree and chIp Sos short time exposure. Dry Stare Setts In general, the hi gher the frequency, the lower the displace- Figure A-7. Coasttng Noise on Various sent smpiitsde must be it saintain a given level of comfort. A Surfaces. 7.5 Metres, Average of Three Vehicles [126] plot of subject ive responses no 'ibration as a function of dis- psasement amplitude and 'requency is shown in [130). Whmie the results dIscussed asove represent an attempt to de- rine vibration toleronce limits, it would be wrong to assume that discomfort is soieiy related to the body's dynamic response to vibration, or that any simple nsi'aericsl expressior of human toler- car. be universa lay applied. As in the ease of annoyance by noise, disturbance by vibration of any given intensity will depend mind I vidua 1 susc ept mit sty and the parti t ui ar circumstances. In

A-47 A-aa 22

a home, for example, the limit of acceptability or vibration may both objectively and subjectively. The former consists of placing lie on, or a little above, the threehold of perception (on the accelorosetefa at selected positions within a vehicle and recor- order of 0.02 m/sec2), whereas in public transport a level of some ding accelerations while traveling over selected roads. The lat- 20 db higher than this may be accepted as not uncomfortable. ter consists of weighing the opinions of individual subjects in their reaction to vehicle and road characteristics. Objective In general, the human body is less able to tolerate vibration measures sill be used in the present dissuesion. than is the vehicle structure. Thus, vibrations transmitted through the vehicle body, although having little effect on the Comparisons between the acceleration spectra of 28 different vehicle, may be somewhat more detrimental to the driver. vehicles as measured by Oliver and 5-mitehead [128], and the ISO prop osed standards discussed in [121] show that the indicated vi- Specific detrimental effects that have been noted relative to bration levels in all vehicles except one are below epatigue_ exposure to prolonged vibration include [1321: cecressed Proficiency' (FOP) limits for up to eight hours of ex- Degraded visual acuity as manifested in blurring of the posure. The vehicle not meeting the eight-hour exposure limit (a visual image on the retina- (The result of relative light commercial vsn) would, according to the proposed ISO stan- movement between the eye and the viewed object.) dards, cause some POP after about five hours (an ae,plitude of 0.059 Degraded reaction time in performing simple decision in the 2-4 and 4-8 he bands). tasks - The effects of road surface and speed on vehicle spectra are Degraded performance in tracking taska--tesks which make illustrated by data collected by chiesa and Oberto [118]. These up a large part of the driving tamk. data are vibration spectra for three automobiles (s set which ex- hibited the highest amplitude spectra out of a total of six vehi- Road-induced vibration can, therefore, have serious effects on cles tested), that were tested on three road surfaces (uneven, driving performance. The effects are, however, dependent upon cobbled, smooth) at three different speeds (40, 65, and 90 kmVhr). vibration frequency, intensity, and duration of exposure. All measurement a were made with accelerometers fixed to the front A.6.2.2 Vehicle/Roadway Vibration Characteristics. Road part of the vehicle body at a rigid point on the frame above the surface-induced vehicle vibrations which could result from damage front wheels, (These mountings should produce vibrations of by studded tires are in the general subject area of vehicle ride. greeter assplitude than those produced from seat-mounted instru- The vibration path which is pertinent to ride conaiderationa ments.) The vehicles used weighed between 2.600 and 3,100 lb. (i.e., the path associated with whole body vibration) includes, in when loaded with three passengers and instrumentation. order, the road, the tires, the suspension, the vehicle body, and - In examining the various spectra of chiese and Oberto it is the seat. Each of these produces an attenuating effect on the evident that there is a general increase in vibration aiicplitudem road surface forcing function, such that the vibrations felt by with speed. It can also be noted that amplitudes increase roughly the passenger are substantially different than those occurring in in the order of the smooth, cobbled, and uneven surfaces. With the road. Further, individual differences in the aforementioned few exceptions, however, maximum accelerations are less than 0.5 vehicle components can produce an order of magnitude difference in m,'sec2. Vibration induced FOP should not be expected, then, for passenger-sensed effects while traveling over the same road surface. at least four to six hours (see Hanes [1231). Vehicle ride (vibration) characteriatics are commonly measured

A-50 A-49

A.6.2.3 Vibration Produced by Pavemen:s Roughened by Studded deleterious- No combined tolerance limits for both stimuli have Tires. The spectral data presented by chiesa and Oberto [118] un- been uncovered, however. doubtedly encompaas the road surface conditions which can result from studded tire damage- comparison of these spectra with vibra- A.? EflCTED STUDS THROWN FRON NIGH-SPEED VEHICLES tion tolerance standards indicate that an exposure of at least Experimental evidence has shown that studs are lost from four hours would be needed to produce Fatigue-Decreaeed proficien- studded tires during operation usage. Experience has shown that cy (POP). studs are more frequently lost from stud rows which are nearer Dare should be taken in interpreting this result, however, the center of the tread pattern [138, 140]. other evidence for as Van Deusen [135] has pointed out, it is difficult to inter- clearly indicates that studs are lost much more frequently during panic stopping and starting maneuvers, than during ordinary pret random vibration data (which contains all, frequencies) in terms that can be compared with tolerance limits. While it might driving [138]. fIre wear appears to be an additional factor af- appear that a vibration becomes limiting whenever a measured in- fecting increased loss rates [1421. While It is clear that stud tensity in any one frequency band exceeds a limiting value, this loss can be expected as the result or studded tire use, loss procedure ignores the additive effects of adjacent frequency bands. rates cannot be easily quantified. In addition to the factors mentioned above, other factors affecting stud loss include stud It seem,a reasonable to assume that the human does not respond independently to each frequency component, but that his responae is type, stud protrusion, stud wear, tire compound, tread pattern, stud hold shape (in tread), stud installation pressure, and the a cumulative effect representing all frequencies. Since vibration tolerance data is determined from subjective response at a single thickness of the undertread between the stud flange and tire frequency, it is not surprising that attempts to interpret these cords. Whatever the factors involved, however, it is clear that results in terms of vehicle ride comfort have not been completely studs are lost. successful, in the complicated vehicle vibration environment, - The operational loss, or ejection of studs has been reported then, vibration tolerance data should only be used as a rough to be frequent but not hacardous. In the early years of studded guide. tire use in Nichigan there were a rash of reports of police cars In light of these comments, FOP due to travel over studded being shot at during freeway patrol. Evidence seems to suggest tire-damaged surfaces may very well occur at time intervals sub- that most of the events were the result of ejected studs. Such stantially less than the four hours mentioned earlier. reports have become relatively uncommon in more recent years, however. This implies that (1) the source of the problem has been Ad.] SUI,U'IARY. Roth noise and vibration effects, resulting identified, or (2) stud fastening methods have improved. from travel over roade which have been worn by studded tires, have been found to have an effect on driver fatigue ,but only after long The mechanism of stud loss is generally due to sliding be- term exposure. Neither effect, by itself, is considered to be a tween the tire and road surface. Research in Sweden has shown serious factor in accident causation, however, since the time in- that for studs to be ejected by centrifugal force, the vehicle tervals required to produce SD? are generally longer than most speed must be 500 mph [141]. Lost studs, however, often appear people drive without resting. Fatigue effects resulting from a during low speed travel, frequently in low gear, and generally pos- combined noise and vibration environment are undoubtedly more sess very little kinetic energy [139]. Nevertheless, the Italian A52 A-51 23

jovernmert nas :'ecoomenuedt-: flaps be fitted behind rear wheels when loon art-used. Table .4-9 Stuc ElectS op Velocitiel Toe ns,zsr.i nvclve.: from a ',,,at or ejected stud is illustrated on Figure A-b. Four points are shown which will be assumed to be post stone of poys it In r.tud eloctios along the tire circumference. EjectIon Pont Ejection Velocity if it illumed silt s.c hr's is rolltng without slip (a good 1 0 Sr se;tpt :cnrorhlt:-tpae d travel), then the stud ejection veloci - 2 V (1-cos +) ties for 5:-to various poInt s'-a a" indicated on Table A-9. Clearly, 3 V C's arbor e rt,idl des ted at -a,, log las (between point e 1 and 2) 44 2V will oScar the roar ova-han' of 'e vehicle and become a hazard to oscossir-' traffic [143]. It is clear, alto, that the backward ejec- 'The vehicle Sc aln,,med to he traveling at velocity V. .:,rityofsucissvudr will he relatively low. F or values of oat;,- c, C 250 , the backward ejection velocity will be less In ounmary, etud loss, althouoh occurring rather frequently-- IrIs: lii of the vehcle velocIty. (Values of + for con temporary ospeciaily as tires become worn--should not be considered a major aurosottlo-a'erc between 8 and 2 44" 1254].) Under these conditions, fattori naccidantcaucaticn. the stud will be a hazard primarily as the result of the velocity of a foil owing vehicle. The dsnger is sisilar to that of an ejec- A. I VEHICLE COMPONENT DEGRADATION ted stone of sial lar cads. VehIcle component -legradat ion acehmnisss are s function of aging and loading procelses. The former consist of corrosion, rot, sunlight, 'let erlorat son, 'tc., while the latter Include fatigue failure, airmaive wear • buckling, creep, and fracture [149]. Road surface:; dsmagcd by s tudcied tires may affect load cycling processes, i,ut ha ye no a ffect on aging. The components which will be most at, affec ted are those wi,ici. mike up the steering and suspension systems

The primary otuddeu tire-produce d read surface asperities which could cuuso adverse component loading conditions are the longitud.inal ruts. Och,,r ssprrit ies, such as esposed aggregate, are of an orcjer of magnItude in sIze which is easily engulfed by the tire tread. Although noisy, component cycling is not Involved. l.onssl tud Ins 5 ruts • nopeci ally thos e with steep edges, can '1(1-ce- $) product sIgnificant steering lystem inputs (see Section A.4). Traveling from sid.-to-s iue within arut could, therefore, increase Figuro A-B. Stud Eject ion Oeometry the cycling times of steering cyst em components. In addition, such

A-53 A-54 travel weul'} als:o incri,-,-:, ti,ticllr:,; i,itorles of suspension are isssediaeely evident: component::. Dofleet inn 'smplitud'et would be Sow, however. The The number of vehicle models, the number of individual fortes and vibrat ions witch are neco:n.ary to produce long term components, and the required test mileage are so large as degradation would undsu:.t',diy he sensed hyadrivor as annoying, to make a full scale test program cospletely impractical. however. A driver, in -.urn, could in ospected to steer out of and Crow': I to one 11 do a fc uci, ruts, th'so ml imi rating the distur- Data collected for one set of vehicles would become obso- bance. At: nol.od In section A.14, however, no tendency for steer- lete as soon as the results of the data are used to modi- ing out of rut:; (set yet teen identified. fy components and change their failure characteristics.

If it is aroused ,soa'-ver, that ruts of significant depth The effects of secondary fact ors such as proper lubrica- t',therd,sexm br will 'c s'st ho some future time, the question tion, corrosion envirsnsent, and minor changes instruc- tircn It-some:, what wi 111', the espected component degradation re- tural strength could be large to fatigue failure statistics. from such ruts? unfortunately, very little information 1:: avoilat le which ca's load toa satisfactory answer. Evidence By and large, then, the systematic collection of data which could further s usse':to th,,t roll 5 at i e random load fatigue testing, be used In drawing meaningful cone lueisns about component wear - utmilar t, l,;s,l in,' conditions 'iuring actual highway travel, is and failure rates is cons idered to be impractical, i.e., data ult now irs; ng incorporeted into automobile component design concerning wear and failure rates which result from travel over ,standurds [148, 1511. Data that are available on actual component pavements roughned by studded tIre use. reliability are, for the nost part, in private hands and not avail- mven if data were available that indicated substantial com- able to the puttss [152]. ponent wear and fatigue rates, however, the effects on vehicle The only psilished data fesundkrss that of a few fragmentary safety and subsequent scold ent caunetion are problematical. Bird test,; which were r.ondur.-,.ed at the Motor Industry Research Associa- at al. [145] and Fancier, at al. [147] have shown that most suspen- tion's (5(15(A) pave' (cobbled) track In England [146]. One thousand sion and steering components must be worn considerably before any ef- cUes on the piv6 l:rackts c oosi dereci to be equivalent to the feet on vehicle performance in apparent. Asynops ii of this work is lifntime of a vehicle or normal roa do. The subject tests on on- given by PIncher, at al. [147] and can be reviewed there by the reader specifirsd vehicles prods;ced the foilowing component failures: who is interested in gr'sater detail. Briefly, the conclusions reached relative to varIous kinds of steering and suspension com- I. An, idler arm s.u:,port treeket iroken after 29 basis miles. ponents are listed as fellows; 2. A front ousponslon crosa cerher cracked after 850 pavf miles. Shock Absorbers - Shock absorber degradation contributes to '5. A steering arm hr'sken after 912 pave miles. a deterioration in vehicle handling performance under conditions ,shdch excite vertical vltrstions. The phenomena which result are on a statIstical basis these data indicate little more than menifcetatiorsof "braka hop', 'wheel hop', and 'axle tramp'. the fact thot oh, nvent:- happened, and are of no value in arriving The:;0 at. rent rd icssd ce,,eluo Ion... phenomena contribute to a net loss in tire shear force and roduce beth braking and currn.ring performance. Ever with vastly In cc,r ,l tuening trer quns slop of gather iog the needed data to dovr:sded shock absorbers, however, changes in performance are not errivi, at :;latioiically ::iroifscsnc answers, the following problems large. Removing 90% of the fluid from shoek sbsoriers was found

A55 A-Se 24

to cause premature wheel lock-up in braking maneuvers and a loss effects were apparent, however. of roadhoidtng capability in a turn [147). The loss in braking In su,msary, the road surface asperities resulting from performance was on the order of 20%. In this same test series, studded tire damage which could cause adverse component wear rates however, using shock absorb era with worn guide rods and broken are the longitudinei wheel ruts. Traveling from side- to-side on valve guides (coesison manifestations of extreme wear) produced no such ruts will increase the wear rates on steering and suapension perceptibie change in roadholding performance. components. No relevant wear rate data are available, however, Other work has shown that shock absorber damping characteris- which could be used to define the severity of the problem. In any tics mumt be degraded at least 50% from nominal before signiricant case, even if wear rates are substantial, the effects of component differences in vehicle performance or driver skill requirements degradation on accident causation have been found to be of little are apparent [liii, 145]. It was also found that a 50% decrease consequence unless the wear is of such magnitude as to virtually In damp ing reduced maximum lateral acceleration capability (tur- require component replacement. ning performance) by 25 to 40%. Again wheel hop was the major factor. Ball joints - Loose ball joints have little effect on vehicle handling performance, short of complete failure. Steering System Play - mtmering wheel play seems to have little effect on path following accuracy until play at the steering wheel periphery eaceeda 3.0 to 3.5 in. [145]. The main effect of steering wheel play seems to be an increase in driver work load to account for increased peak-to-pea k steering wheel motion. A miailar conclusion was reported by Panther, at ml. (147), where it was found that, 5lndeterminancies of front wheel steer angle which arise due to lash in wheel bearings, ball joints, tie-rod ends, or in the steering gear box, whether taken singly or in combination, do not exhibit a first-order influence on vehicle lieit performance." Other components- Front wheel misalignment was found to be an insignificant factor be path following accuracy for all but 'he most severe maneuvers [145. 147]. Vehicle rollover resistance may, for example, be influenced by front end misalignment.

The bce of 5 stabilicer bar and a steering damper have also been investigated by mird, et.al.[145]. Each was found to alter vehicle handling characteristics and in some cases increased the amount of control activity required. No mediate safety-related

A-SB A-57

APPENDIX b IN-SERVICE PREDICTION OF RYOHOPLAHINO AND WET SKID

Dynamic hydroplaning and wet a kid are considered to be poten- tially the most serious driving hecards resulting from the pavement wear patterns produced by studded tires. Accordingly, a nomograph

has been constructed to aid in sseeeeing the relative aeneitivi- ties of the variables of influence, and to orovide quick estimates of the aoeeds at which dynamic hydronlaning and wet skid might take elate. The results obtained with the nomomraoh have not

been checked against actual emnerience, and it cannot be offered at this time as a tooi of oractical usefulness. However, it 11-

lustrs tea a means that furt her study may show to have very practical apolication in determining when danr.eroue situst ions are

developing and when improvements should be mmdc. The nomograch isahowr. in Figure 3-1, The details of how the nomoeraph was developed and how it can be used are described in this sopendix.

B. 1 HOMOGRAPH DEVSLOPOtiNT The nomograph ia based on the application of Equations (A-5) and (A-i) of Appendix A. Equation (a-i) [119] has been modified to include the effect of water accumulating in sworn wheel path rut and is rewritt.ui as folbows

d - [3.38 5 10_3(l/T)_l1(L)43(1)59(1/S)42] - T + H (3-1)

wherc each of the parameters is listed in Table B-i slong with its definition and in-se rvice range of values. The ranges are self- explanatory, except, perhaps, for texture depth, rainfall inteneity, and truss-slope.

s-a A B 9 A'

[2 0.05

L 1 H L0.10 ! H I VP --I I 0- 15 -4 "0 Ce 0 04' 0.10 I -1 - 0 L-.2 .c a 4-i p I Ce F-0.20 ° I 41! 0) 0 c-o. is -- a-i-L -i CD I -. -0.25

-0.20 F- 0.4

0.30 r3 I 'VI L025 0.5 (Find A Value on A' Scale) -0.35

Tire Lift Coefficient (Cu) 30 2.0 t A. conventional SM4PLE SOLUTTON 1. snooth 100 2. long, rib tread (1/4 itt) 26 S = 48 3. closed tread (1/4 in) 18 I = 0.8 in./hr /o 60 L B' 4. open tread (1/4 in) 15 = 20 it B. Radial T = 0.10 in. smooth R = 0.20 in. 60 = 25 closed tread 25 C 0 .01 K = 0.65 / 50 \ s I open tread 16 H 0 'I s' Solution = 24 mph / 'I / cf-i C Pavement Surface Fe ctor 60 40 (K) 0 70 It Smooth 0.15 0 -4 0.02 te Fine textured, rounded 0.38 tso 4' 10• Fine textured, gritty 0 ' U 1.00 40 16 \ a Coarse textured, rounded 0.65 L 70 4 30 i-i Coarse textured, gritty 2.90 a 0 o I ,122 8 a \i\ C 14 0.05 / -4 a)I aC [80 E20 - \ o 2 0 0.10 Ii 0 NOTE: FOR DEMONSTRATION ONLY 0.20 tO.30 90 10 F 0.50 (Find B Value Figure B-i. Hydroplaning or wet-skid prediction nomograph. on B Scale)

In examining Equation (8-1), the term in brackets on the eight TABLE 8-1 ide represents the run-off rate, the P term is used to correct the WATER DEPTH VARIABLES depth levels to the tops of the asperities, and the H term represents water accumulation in ruts, Although it was recognized that Variable Definition Units In-Service Range the pavemenl crown in most initancei will produce a situation in T Surface Texture Depth in. 0-0.25 In. which the water-entrsnrnent deoth will, be lsam than the rut depth, I, Drainage Path Length ft. 12-60 ft. the two were assumed to be equal in the develonment of the nomo- Rainfall Intensity in/hr 0-5 in./hr graph. Little or no informatior is available on the differences that actually exist, and the amsumotion is conaervative. If the S Pavement cross-Slope ft/ft 0 to 1/8 ft/ft H Rut Depth In. 0 to 0.5 in. water level is below the loom of the asperities, d can be negative. As will become evident in using the nomograph, the dominant term is the rut depth. This is true for rut denthl sbove 0.2 inches Texture depth foj root pavement a rarely averagea; more than and for all reasor.alsle values of the other narsmeteri. 0.10 inches [119. 240]. The indicated range was excended to 0.25 Equation (A-5) is rearitLen as follows: inches to account for a6gregste exposure which may result from studded tire wear. The indicated maximum cross-slope IC the level Vu = 7,61 which is considered by the AASHO [237] to be the maximum practical To C N (8-2) superelevation rate for curve design (i.e., e a 0.12). where Rainfall intensity and typical yearly duration experience are am - wheel load (18) gi veil in Tablo B-2. Since duratIon experience vsrlea widely from 8 = tire tread width (in.) coo geographical region to another, the data showm represent little - tire lift coefficient more than a rough indicator. In general, though, rainfall intensi- water depth (in.) ty greater than 1.0 in/hr is very uncommon. Rainfall of such severity would undoubtedly require a driver to alow down for lack of This equation is supported by a considerable amount of experimen- visibility, regardless of the hydroplaning threst. tal data which were obtained from a rotating drum tire testing machine [11. 121. In the experiments, the drum is initially rota- TABLE 8-2 ted at a speed which is somewhat above the hydroplaning lift-off RAINFALL INTENSITY EXPERIENCE [312, 641 speed. At this speed the cc at tire is not rotating. To determine Vu• the drum is gradually slowed until the tire starts to rotate. Typical Yearly Duration (hr/yr) Name Intenaity (tn/hr) The speed at which rotation begins, then, is called the spin-up speed, As noted in Section A.i of Appendix A, Vu is lower Driccie 0.01 32 Vu than the spin-down speed, Vd, which is cononly associated with Light Rain 0.04 - 0.20 20 automobile tire hydroplaning. Vu therefore, represents a more Hoavy Rain 0,50 - 0.80 13 Heavy Downpour 1.0 1.1 Very heavy Storm 4.0 0,4

5-4 5-3

conservative approach to predioting incipient hydroplaning or wet ,his would be called a fine textured, gritty surface. In order slid. The validity of Equation (1-2) had been checked against in- to determine the effect of surfaro properties on the hydroplaning dependently derived data [611] and Had indeed been round to Be con- or wet skid speed, Vu as computed from Equation (8-2), was multi- servati ye. (The independent data was ohtsine.i from hydroplaning plied by a surface factor constant. For a fine textured, gritty teats which utilized a sowed trailer and a toot trough with con- surface the constant is 1.0. For other surfaces, the constant trolled water depth.) 'as determined by the followioc ratio: In using Equation [B-2), valuer of the ratio W/B were obtained V.,(surface n) (8-4) -or typical vehicles [2)8, 2591 and values of 0 for typical tires K = V(fine textured, gritty surface) [11, 121. - W/Bvaluea were found to range hetween 68 and 2110 lb/in. The where VH is the hydroplaning, or wet skid speed. Vmlues of K smal 1cr values are associated with small automoBiles of the were computed from two separate acts of consistent data (i.e., muse tires, tire pressure, tire loadog, and water depth) and foreign car' class (weight 1.500 to 2,000 ib), whereas the larger 2, 281. These values 'are shown on the nomograph. vo li.es i f typical of 5: all oil wagc-ra and limilullre 1111, automo- averaged 1 bile,'. This isanintercitimp, finding jainre it Implies that ama sir,' vehicirj are moe ii ke]s ta" spa's'] clue hyalb'op laming. In 8.2 NONO0RAPH USE corstract.lng the nomorrsaih a value ,,f iy lh,'i n. woe used for v/B. The urper row of t:ae nomograrh of Figure 1-1 represents a This xoans that the hydroplaning or wet skid apee'd determined grsrhicsl solution of Equation (1-1). The lower row represents from tilt nosiographi • will haeauortt case condition. Equation (5-c). A same le solution is shown in the figure.

Values of 011 for tjpical tires are shown or the nomogrsph. As indicated in Appendix A, eonr5rr1sch found f,lr smooth, conventional tiroi that c.1 varied aocordl jig to the following formula'

l.]tl , 01] - I5 t,, ) (B-))

where W = prescribed maximum tire local capacity

In practice, W0/w is no sore than about I.B. Therefore, a worst case value of 0H for aattooth conventional tire is about 100. This is the value ihown on the ns,mograph. Surface effects were diffIcult to dote nine since all of the oengenbaoh data was collected using the same eurfaco--a water- proof abrasive lining of grain size 80. (This eeanx that the surface abrasive must paaa through a lacreen of Ba meshes/lineal inch.) under Kusuer's c1ac:ificatioe aystem (see Section A.1.2) 27

APPENDIX C The most attractive locattcn for the investigation is one of PROPOSED RESEARCH PLAN the turnpike systems located in the northern state... Those systems possess several features which strongly recommend tI'ra. The, Empirical verification of t]ce advc'rnce cci rety effects of present the high-volume, well -coast ructnd rod dr, tui;i I. w III 'ep;',' - studded tire- in,ii,ceil reek I datc,ir, precirkil;; mclii a cliff Icu Itirs. The ence sub- t ant to I studded tire damage. The traffIc; nec ace Iclent problems arise from the wide variety of precnnditi 'na which may data available from them is good. Norma 1 accounting procedures play a role in any particular group of accidents and from the ensure accurate counts of traffic volume. The roads are well pa- rather poorly understood mechanisms which produce a crash under a trolled by specially assigned polIce uni ta which will yield quite particular set of ctrcucstances. The work necessary to directly full accident cover age and a high standard of report quality. verify the effects of studded tire damage on crashes will be cx- Administratively, the turnpikes are operated by a aesl-aut onomous tenslvsc, and the decision to embark on cuchaproject must weigh state agency which will mintm'se the problems of operating a carefully the substantial likelihood of Inconclusive results in multi-phase investigation'. this area against the potential payoffs from other research; e.g., Yet, the turnpikes' unique advantages pose problems of their the nced for better definition of the properties of the tire/road nwn. First, site possibly for reason- indicated above, the Interface. turnpikes have an accident experience which is far better than An experimental research plan should involve four major In- the typical portions of the Interstate system. The generality of terrelated activities; (1) definition of the extent of damage to turnpike resu its might be questioned, and detecting differences a particular road network, (2) examination of the presently avai- on 5 lower accident rate base imposes more stringent statistical lable accident data to detect any assncistion between studded requirescnts . Further, the uniformly high maintenance standards tire damage and accident patterns (3) observations of traffic of tire turnpikes may prevent as severe damage from appearing as flow behavior to ascertain differences between lightly damaged would be found on some Interstates. Both of these problems sug- and heavily damaged roadway sections. and (II) collection of sup- gest the possibility of a Type 15 error in not finding an effect plesental information from police accident investigations to deter- when, indeed, one is present. In this sense the present program sine unique patterns of vehicle dynamics in the crash sequence poses a strong test of the adverse safety effects of atud damage, associated with studded tire-damaged roadway sections. since if an effect is found on ares cgnizably safe road system, it The investigation would be conducted in two phases. DurIng constitutes good evid ense of hazards elsowhare. the first phase, measurassnt of studded tire damage and analysis Problems that were likely to crop up in using regular Inter- of accident data will be pursued. This effort will, show whether state or other limited-access highways also weighed heavily in studded tire damage is sufficiently extensive to present a wide- the recosssendition for using toll roads. First, to detect pomsiily spread potential hazard, and if the potential exists, whether or not the damaged sections disolay an accident experience that dif- 'The decis ion to use turnpike authorities is not intended as any fers substantially from that of the undamaged sections. If both reflection on the excellen t research efforts and facilities of state highway departments, but the turnpikes are in the unique po- proccaitions hold true, than the next phase, involving activities sition of having central responsibility for all phases of opersticn three and four sari be teolesented to demonstrate more clearly the including construction, maintenance, traffic, and enforcement which greatly simplifies the problems of organizing the research precise natures f the hazards treated by the studded tire damage. project.

c-I fine statistical differences will require highly aceur ate traffic 0.1 SURVEY OP ROAD DAMSON volume information, which would require an empensive, additional Before any .'thermeamingfws Pea earch can begIn, a survey of research investment on non-toll roads. second, traffic conditions road damage Is needed, both to alsow an aasosseent of the relative on turnpikes tend to be more uniform through the absence of heavy severity of the potential hazard and to determi,,e if sufficiently coeesnuter traffic and frequent interchanges found in the urban por- large sections can be found to make statistically valid compari- tions of the Interstate system. Third, although little hard data sons of accident experience. While the potential for damage has is available to substantiate the assumption, the typical turnpike been exparimentally verified and has been of great concern, areas driver may be both more mature and literally more sober than the of severe damage presently have been con fined to limited highway typical freeway motorist. All these points suggest, then, that sections where conditions have been favorable for its development. freeways will tend to have more mxtran e ous variation in their data As indicated in the anaLysis of accident experience which follows, and hence the turnpikes' advantages would seem to outweigh their determinstion of the differences in accident potential requires problems. large roadway segments, but our informs 1 ssmnning of several ex- In formulating this experimental plan, officials of three tensive highway stretches indicated only slight damage to most turnpikes, the New York Thruway, the Ohio Turnpike, and the India- areaa. Finally as a matter of analytical necessity, it must be na Toil Road, were contacted to discuss the anticipated research possible to link accident data to the characteristics of particu- effort. All three agencies expressed interest in possible parti- lar road segments. Hense there is a need for systematic assessment cipation, provided that appropriate arrangements were made, and of the damage before proceedings. offered many useful, concrete suggestions on the operational pro- Two types of stud-Induced changes must be aesesaed; the extent blems of conducting the program. mmmcd on the observations during of rutting, and alterations in pavement friction properties. the visits, the New York Thruway has mxperiensed the most severe While methods of characterizing the pavement on both these dimen- damage and possesses a sufficiently large traffic volume to offer sions are still evolving, presently available techniques will be good experimental possibilities. The Ohio Turnpike has experienced used to collect data. l5or changes in pavement friction properties, somewhat less damage than the New York Thruway, but has already the usual skid trail er teci.ntques can be used. Measurement of coepletad the first program element (a survey of studded tire da- wheel path wear is not standardized, but a taohnique similar to mage) and of the three .iystess, Ohio seemed particularly receptive Xeyser'a [166] seems most procising. The amount of data neoeiaary to research efforts. The Indiana Toll Road fortunately has es- for analysis will depend on the inherent variation in wear pat- caped the extent of damage suffered by the other two highways and terns, and m two-stage sampling plan is suggested. Both issues are therefore ii not a likely candidate. letween New York and Ohio, discussed in detail below. the choice is close; final determination should be made only after Ccncern with the potential damage Is still legitimate, since the more fully exploring the program with these aganciws. limited damage to date' probably portends whmtwill eventually occur elsewhere. It is quite possible, though, that the problem cur- In the sections that follow each task will be discussed in rently is not researchable at the level of determining differences more detail. The experimental design estimates which have been in accident rates, since sufficiently great segments of highway prepared are based on the New York data, but can be applied to have not yet elcperiensed damage. term damage is not used with respect to changes in the pave- Ohio with slight modification. ment friction charactertstics, since the action of studded tires may or may not adversely affect these properties.

0-3 0-4 28

0.1.1 MEASUREMENT OP WHEEL PATH WEAR. Wheel path wear mea- available computer facilities to convert the tape records to digi- surement by itself is an inherently simple p,..Co.. All one needs tal form. The set-up costs for the electronic recording will be to collect quite aocurate data is a long, reasonably firm straight higher than for paper and pen, but the operating costs per obser- edge and a machinists! square. The problem is to collect rapidiy vatton will be lower. large amounts of data under traffic conditions to minimize traffic Preliminary comparative cost analysis of the two technique, congestion and crew hazard. While more sophisticated means might is illustrated in Table C-I. For 500 observations, the mechani- be devised. Keyser?a proftuometer [166] modified to cover the en- cal technique has a clear cost advantage over electronic recor- tire twelve-foot lane width and mounted on a truck's hydraulic ding due to high initial equipment cost for the more sophisticated tailgate, appears to be the most convenient approach. The recor- approach. For 2500 observations, the costs are roughly comparable ding medium used can be either paper strips, as originally used by if the on-road collection rate for electronic means allows a 20% Keyser, or an electronic system. For paper strips, a lever arm fast er recording rate. If data recording occurs at the same rate resting at one end on the surface and attached to a tracing pen for both techniques, then the point at which the electronic ap- will produce an accurate record of the profile. The records can proach breaka even is in the 14.000-16,000 range. Equipment cost then be hand-measured at suitable intervals (say three inchee), estimates were based on the informed opinion of HSRI technical the measurements tabulated, and these tabulations entered onto personnel, and included machining and installation of mechanical punch cards and into the coaputer. This approach requires a mini- coaponenta, purchase and assembly of unique electronic componenta, mum of specialized equipment and personnel and, conaequently, will and rental of a data recorder. Coliecti on costs were estimated need only a short time to implement. However, the data reduction on the basis of a two technic ian crew obtaining four observations process, particularly for large amounts of information, can be per hour uaing pen and paper techniquea and five per hour with time conmuaing. Alternatively, electronic technique, can be used electronic means*. H and measuring and keypunching costs are baaed to record the data. Instead of a pen tracing on paper, the level on performance of similar coding tasks, and the computer processing arm would be attached to a potentiometer (pot) and the vertical costs were estimated from the current charge structure of the HSHr displacement would be recorded as a signal on one track of a tape 1800/AD4 hybrid compuis. facility. Since actual costs will vary data recorder. Similarly, the laterai distance across the road from organization to organization, a more detailed analysis will be can be recorded with a atring pot, and the degree of super-eleva- desirable before making the final determination of the approach to tion with a pendulum pot, on the second and third tracks respec- be used. A conaervative approach would use pen recordings in the tively. The fourth track can be used to mark particular sites. first atage sample, and would then use the esperience and the in- cnce recorded, the tapes can be processed with an analog to digi- dicated sample size to determine if electronic recording means tal converter, and the data will be ready for analysis. This will be cost-effective. method bypasses hand processing, which introduc es errors, both in the field in positioning the paper record strips, and in the office in measuring and tabulating. Pield operation may also be alightly faster by eliminating steps necessary to change paper 'auth estimates may be somewhat optimistic,s Inc.they both have strips and to adjust pens for each observation. As a disadvantage, assumed eight minutes of travel time between sites and a single electronic recording requires instrumentation engineers and tech- unit operation. A second vehicle with driversmay benecessary for traffic protection which would increase the costs by aboot 30%. nician, to design and to asseable the equipment, and lea, widely c-s c-i

Figure 0-1 where the first segment is the entire road between interchange 1 and interchange 2, the second segment is the portS on Table C-i of the road between a,iterchamge 2 and the end of type 1 pavement, Preliminary comparative Cost gatimates of observation of Wheel Path Wear and the third segment is the balanre of type 2;avement between interchange 2 and Interchange 3. The first stage effort will be pen & Paper Recording Electronic Recording to define the degree of variance within a particular segment, and H,mber of Observations Number of Observations the second stage effort will be to ecamine wear patterns among 2500 Item 500 2500 500 segments.' A basic research hypothesis is that the wear found within a Equipment $ 500 $ 500 $3500 $3500 pavement segment is relatively uniform. The suggested wear mea- Field Crew $1875 $9375 $1500 $7500 surement program will test this hypothesis. The initial detailed Tabulation $ 200 $1000 $ -- $ -- measurements will establish within-segment variation and will de- Keypunching $ 70 $ 350 .$ -- $ termine the number of measurements needed to establish among-seg- Computer cpa. $ 30 $ 150 $ 200 $1000 ment differences in wear. If the wear found in a segment is both TOTALE $2675 $11375 $5200 $12000 uniform and pervasive, then it can be assumed that the damage will affect traffic over the entire segment, and that meaningful sta- c.1.2 STATISTICAL VALI0ITY OF WEAR MEASUREMENTS. A two-stage tistical comparisons can be made about accident esperience among procedure will be used to describe stud-and uced changes in pavement segments. characteristics. The first stag ewill measure wear patterns over The suggested research plan has be en construe ted on the assum- two relatively short intervals of highway, one ltghtly damaged and ption that wear will be uniform over relatively long segments. one heavily damaged, to determine the variance in wear patterns for The uniformity assumption, if valid, permits the program to be a particular highway segment. If wear patterns are fairly uniform, confined to a manageible number of measurements and allows then a limited number of measurements w ill be taken from each straightforward testing of statistical hypotheses. If, however, highway segment to characterize the wear pattern over the entire the uniformity assumption does not hold, the research problems will highway system. become far more substantial--possibly so substantial that no mean- As dtacus,ed earlier, the amount of stud-Induced wear is a Ingful reaults could be estractsd. Should, for esample, the wear functi on of volume of vehicles with ste.dded tires, traffic move- patterns be such that only short segments of a few yards to around menta, and highway materials. Of theme three factors, traffic vo- a mile have relatively uniform damage, then the propoaedresearch lume and highway materials are relevant to wear patterns on main design would be invalid. Under conditions of cmi reme variability, lanes of high-volume, inter,tate type roads. Within this contest, it indeed would be difficult to perform any empirical research. a highway segment is defined as the length of pavement over which 'One slight problem with this definition is the treatment of portion, of the roadway within an interchange area such as between the traffic volume and construction materials used are constant; ramps. Two approaches can be used. One is to arbitrarily define a segment is either the entire road between two interchanges, or the boundary between the two segments as the point midway between the two most widely separated ramps. A more analytically conven- the portion of an inter-interchange segment which was constructed ient approach is to consider the interchange area and some distance with the came type of materials. The definition is illustrated in from it, say .25 mile, as a distinct pavement segment. 29

About the beat that could be done would be to take intensive wear measurement for a large sample of segments, and to obtain similar measurements at a large number of accident sites. The reaearcherr, could then compare the emtent of damage at accident sites with that at sample sites, in hopes of inferring inductively more severe damage being assocIated with accident locations. However, this procedure will be subject to the usual logical pitfalls of any Induct ive study. Unfortunately, no information is currently available to indicate which conditions are likely to hold. Field observations of damage have centered on measuring wear over time at selected spots, rather than on describing damage for long road segments. An investigation of damage similar to the one recomended was in progress in Ohio, but the results are unlikely to be available soon. One of the investigators devoted effort to the problem. His observations left him pessimist it about the ability to research Segment 1 Segment 2 Segment 3 damage patterns in the near future. He, with the assistance of Material ___A______Material Type I Typ e 2 a number of highway officisle, was able to locate only sporadic 000urrenoes of moderate to severe damage. While these areas sight portend future, more extensive damage on more extensive road EWMAIMEZZIM areas, their presently flai ted naturewill pose difficult problems inconduoting any research either of the reconended plan or of Interchange 1 Tntcrohenge 2 Interchange 3 alternative plans. Still, in areas where damage had occurred, it appeared to be fairly uniform over ext ended segments. Hence, the Figure C_n. Segment Definitions recos'msended plan was developed rather than alternative plans. choice of the appropriatc- statistical technique for determining wear uniformity is closely tied to the general problem of characterizing pavements • and, hence, until the more general problem is resolved no entIrely satisfactory statistical approach can be adopted. For the present, the most straightforward approach is to treat the observations as points from a multivariate normal distribution. For each profile measured, the depth of wear for a number of points across the profile can be recorded so that the profile will be described by a vector of observations = (y1, y ..... i. n) there y1 is the depthofwear at the ith point on the profile. 0-9 c-la -

The number of observations across the profile initially needs using conventional techniques similar to those developed by Keyser. to be quite large, so as to yield quite an accurate estimate of The choice of recording moans, me chanical or electronics will de- the structure of intercorrelationa among the points; a measurement pend on a coat-effectiveness analysis and the number of measurements interval of 1.5 or 3. cinches seems appropriate and will produce a per lane per mile needed for statistical validity. Observations vector of 96 or 118 observations per profile. To establish statis- will be divided into two portions. The first set will consist of tical validity, twice as many profiles as observations per profile intensive measurements of a badly worn and of an undamaged pavement should be taken for each segment in the initial sample. Hence, segment. Data collected during the first set will be analyzed to if the original measurement interval was 1.5 inches, 192 profiles determine the number of measurements needed to accurately describe per segment would be required. Once the data has been collected, the wear on a particular pavement segment. If observations can be it can then be analyzed to establish the minimum number of points made less frequently than two per lane per mile, then the second per profile and the sin imue number of profiles per segment neces- set of profiles covering the entire road system can be taken. Si- aary to accurately differentiate between degrees of wear. Without multaneously, skid measurements will be aade at the same points as such a detailed analysis it is not possible to predict the number the profiles. Both sets of data will be reduced to computer usable of profiles per segment since the multivariate approach is suffi- form, and combined with the accident data, described below, in ciently complex to make a 2!i2& judgements unreliable. Although order to determine the relationship between accident occurrence the exact number of measurements per segment cannot be accurately and stud-induced wear patterns. predicted, a reasonable upper bound on the measurements would be two per lane per mile. A higher per mile sampling rate would in- 0.2 ACCIDENT DATA AHAt,YSIS dicate that the wear pattern is so variable that a global approach The second research task is to analyze currently available is not suitable. Further, requiring more than one profile per accident data from the study highway system. The road damage sur- half mile would make data collections coats prohibitive. vey will be linked with the accident data to determine if a sta- Pavement friction will be measured using the conventional tistical meaningful relationship exists between pavement wear and skid trailer techniques. To provide a consistent data base, the crashes. Analysis will be performed for the total number of acci- skid measurements should be conducted at the same point as the la- dents and for seweral subsets of accidents using a variety of teral profiles are taken, and these readings should be incorpora- statistical techniques. A large number of accidents (on the order ted into the same data base as the profile measurements. Both of a year's experience for the Ohio or New York turnpikes) should sets of measurements should be conducted at some standard refer- be used, since it must be conservatively assumed that the relative ence point such as a one-tenth mile marker, so that they can be effect sought is quite small. replicated to show any progressive effects of wear. Two principal difficulties in performing any accident data 0.1.3 SUMMARY. To summarize, pavement wear will be measured analysis are the large inherent variation in crash producing circumstances and the statimtically rare nature of accident event The multivariate normal distribution is discussed in most advanced probabilities. Usually the analyst is confronted with either statistical texts. It is used to deacribe the joint occurrence of trying to find gross multiple order of magnitude effects of a two or more related events. For example, if one were to collect data on the height and weigh t ofanumber of individuals, the pairs countermeasure for a quite limited area, or with trying to deter- of height/weight observations would have a bivariate normal distri- mine dlffuae effects over a wide area. In the gross effects case, bution.

c-il 0-12 30

New York and Ohio systems have had the original portland cement avery apectfic counte rnesasire, such as placing a stop sign at a hi inS corner, can redsce the number of accidents per vehicle mile construction overlaid with asphalt. This might produce a condition in which only relatively short pavement sections of both or passage by many orders of magnitude; but since the particular types dispiay a'psrtloular degree of wear, and consequently after location has a relattvety small number of accidents, the wide allowing for pavement type effects, insufficient data will remain chance fluctuations over time will tend to make statisticslcom- parimons difficult, in the diffuse effects case, a particular to estimate wear effects, in a more extreme case, particular change, such as increassna the number of police traffic patrol degreea of wear may be confined to particular pavement types, thus utterly confounding the effect of pavement type and wear. hours Sn an are a, Wi 11 have only a small effect at any particular Unfortunately, this situation may be quite likely to arise, since locatIon consesuently, , large hue and test areas most be observed for long perious of time to generate reliable statistical the new asphalt overlays will have little wear while the older portiand cement will be more seriously worn. The second problem, test,. However with to:.; obsssrvoton times and large areas, the which Is of less serious consequence, is that driving patterns, underlying assumptIon ti sisilar cnaracterittics between control even apart from differences in traffic volume and vehicle aix, and esperimental areas siesones ir,creaiincly tecuoum . In the Po- llee pa trol example, no two neightornoeds ore quite alike in the can vary from segment to segment of a long highway system. What sole-up of population, traffic rattern, degree of coemerclaliza- effect, if any, this will have on the analysis is problematical, but the interpretation of results must include the consideration tion, and other factors whtth can affect accident experience. As that for both New York and Ohio, certain sections carry large ye- time progresses, travel patterns are likely to c,lter within any lusmea of cos,zsuter traffic while others are oriented predominantly particular areaas nea reads are constructed, new shopping areas to long distance travel'. With the escepiien of the confounding opened, new housing developed, and as other traffic improvement of pavement wear and pavement type effects, the overall problems erogramsoc c:ir. Confronted with such problems, one can either of conducting the accident data analysis are quite manageable. attempt to measure oor.comi tintS ehangas and allow for their effects In the sna lxsls acd fntereretasJsn or reou lts, or one can use Several statistical methods will be used. Initially, pave- multIple IntermedIate ..asur'ss which are more cloaely tied to ment segments will be classified by pavement type and by two or the problem Dole, studied. possibly three wear levels: slightly, moderately, and severely damaged. Standard analysis of variance techniques will be applied Studded tlre-tndu'sed road dlmage appears to be a problem of to find any significant differences in accident ratem among seg- dIffuse effects as defined stove, and consequently measurementi must be taken over a large area fsr an extended period of time. ments aturnpike If st ccc, the effect!; or extraneous changea be- If such differences are found, multiple linear regression texaco test areas can h mInImized to some extent and changes that J. occur, (i.e.. Inornc.s as in police patrol, traffic controa im- 'One possible effect of this problem can be found in New York where the damaged areas are costly in the upstate portions of the provements, ,seathercnnititofli, and fluctuations in traffic volume) Thruway, and the undamaged areas closer to New York City. In can 01 recorded and I ncorporas.ed into the analysis. To some ex- analyzing the relative frequency of wet weather accidents on damaged and undamaged sections, the results may be pertly contamina- tent even ohanges in traffic mIsc between automobile and truck ted by commuter traffic being relatively less sensitive to weather traffic can be controlled throucsh the toll receipt data. Two conditions than long distance trips, since the latter are some- serious problems wI llremair,, however. The first is of a techni- what more discretionary. cal nature. oue to ret-nstructaon, some segments of both the c-14 0-13

techniques will be used to sped fy the relationship among the dangerous than a uniform one. variables more exactly. Specifically, two such relationships that The final analysis activity will exmslne In more dctnil iii'- might be tested are: ferences between wet and dry weather accidents on differing ,nvs- mont segments. Chi-sq,iare analysis of coniingency tables can be ACC1 m a + a1VNILEi + a2AWEARi + a3SOWEANi + aPTYPEi + performed for the following combinations: pavement type by degree of wear by wet and dry weather accidents, pavement type by degree 55PCNWETi + u1 (0-1) of wear by wet and dry weather accidents adjueted for proportion of wet and dry days on segments, and wet weather accidents for pavement type of type of accidents (loss-of-control vs. other 3ln(SDWEA1u1) + ACC1 m a0 + llln(VSSILEi) C a2ln(AwEAR) + a types). The first analysis will test the hypothesis that different segments experience a greater or lesser amount of wet weather (0-2) aFTYPEi + a5ln(PCNWETi) + u accidents depending on the degree of wear experienced, and the second analysis will test to determine if there are significant differences In climate between the worn and unworn segments which where might account for the differences in wet weather accidents. The ArC1 is the number of accidents on the ith segment. third analysis examines the, question of whether differences can be VNILE1 is the number of vehicle miles on the ith segment, attributable to loss-of-control incidents which presumably arise through hydroplmnlng incidenim. An example of the analysis is AWEANJ is the average depth or Wear on the ith segment, presented in Table C-?'. SDWEAR1 is the standard deviation of wear on the ith segment. PTYPE1 is the pavement type, where 0-portland cement and Table 0-2 l=asphali, in the ith segment, Example of contingency Table PCNWET1 is the percent of time that the pavement segment is wet, Pavement Condition ui is a random error term, Pavement Type wet az Total a is the coefficisnt of the jth variable, and is the intercept of the regression line, and Slight Wear X11 X12 X21 X22 in 0 designates the natural logarithm of the particular Moderate Wear variable. X3 Severe Wear X 31 X 32 X Equation (C-i) tests the assumption that there is a strict linear TOTAL X 1 X 2 relationship among the variables, while Equation (0-2) teats the alternative hypothesis that the relationship is curvilinear. All where: the variables are relatively self-explanatory with the possible is number of accidents occurring in the (5, j)th category, exception of SDWEAR, which is included as an attempt to determine .The statistical techniques referenced are described in most intro- if a highly variated wear pattern is relatively more or less ductory statistical texts. c-is c-is 31

X.j or Xi. is the total of the jth column or the ith row and *11W YOlK *5*51 THRUWAY AUS**ORISY X.. is the grand total, II pouaAccnENsers 2 (Xli - iii The X' atatiati = for Ei j being the spec- o ted number of occurrences in each category which E1J = C - ,.---

As indicated earlier, data for the analysis will come from the road damage survey and from standard police accident reports which or are available in digitally coded form from both the New York Thru- way and the Ohio Turnpike, for which a coding form is illuatrated in Figure C-?. Further subdivisions of the data to apecify more complete differences between worn and undamaged segments can be t_ ...... ______— attempted, but it is quite poaeibae that more refined analysis will — = em.,ele ...... not yield statistically significant resuats due to relatively small se.aa.ot..... me.. .waOa,Dse...l,a*tlO a-4'asooC',.m.. amounts of data in particular subdivisions

The amount of data for statistically significant results de- £ lm...a-...... 5-.... a .m,...u.. pends on the type of analysis and on the inherent differences in - accident experience among road segments. The first statistical ap- ,..la.. 5 1fl2 — 1flfl proach--analysis of variance of accident rates among road segments t- wo- -Ze.I i - - --demands the most data. Furthermore, It should be conservatively - -. -. -. assumed that the effects sought are quite small. Hence, the data so specifications are drawn on the basis of an analysis of varianoe of accident rates seeking small differences. Under the appropriate statistical assumptions, the ratio of accident rates per vehicle mile for two segments will have the F distribution so that*: i;ib1hki,!3tt er F( 1/r 2 (c-3) n1) = r - coo...... no'..-. - gTe..-.-, .n5...... where - :rm_ °°' r is the accident rate per vehicle mile (or some convenient nou,...... s, i'-5_..•. multiple thereof),

is the number or accidents used in computing the ith rate.

*The necessary statistical assumptions are that for a given traffic volume, and the number of vehicle miles driven over a segment, the tEf number of accidents follows a Poisson distribution, and that the total vehicle miles on the segment has the Cams distribution. tk; Figure C-?. New York Thruway PolIce Accident Report Form C-lB

The appropriate test is whether the ratio is significantly diffe- road-miles are about equal to i year's experience of the New York rent from 1. In the present case, the ratio of accidents per ve- Thruway for the 5% dl ffc,'once ai,d of the Ohio Turnpike for a 15% hicle mile on damaged and undamaged road segments is to be tested. difference to be statistically uicnifiesnt. Published statistical tables or computer programm indicate signi- arlefly sumaricing, accident data analysis efforts will ficant values of F for combinations of n1 and ri2 . Assuming the merge road damage survey result a with currently available police damaged segment accident rate is 5% higher than the undamaged accident report infcrmatlom. The combined data will be used to rate, 2000 accidents on undamaged and 1500 accidents on damaged study the effects of road damage on overall crash rates and to segments are needed to asko an observed F ratio of 1.05 signifi- determine differences in accident patterns for damaged and undama- cant at the 95% conridence level. To detect a 15% difference, ged sections. Assuming small differences, the amount of data re- 750 undamaged pavement and 500 damaged pavement accidents will be quired by conventional statistical standards will be on the order needed. of a single year's experience for the New York Thruway or Ohio The number of road miles of differing types needed for analy- Turnpike, but the use of such emtenmive systems will run a rela- sis can be dedoced from the numbers above, if the number of acci- tively high risk of contamination by extraneous factors. dents per road-mile is known for each type of pavement. Unfor- tunately, without conducting the pavement damage survey, the pre- 0.3 OETAILED ACCIDENT INVESTr0ATI0N

cise figures cannot be known, but they can be approcimated by If the accident data analysis activity discovers a statisti- taking the average number of accidents over the entire turnpike. cally significant association between wear patterns and accident For the New York Thruway in 1911, there were 9.3 accidents per experience, a further investigation will be needed to verify the main-line road-mile, and from 1966 through June, 1970 the Ohio relationship. Verification is necessary to ensure that observed Turnpike averaged 8.2 accidents per mile. Approximately 80% of effects are not generated by extraneous factors nor by chance. the crashes involved passenger vehicles and non-icy conditions, The recosesended approach is a bi-level accident investigation pro- which reduces the New York crashes per road-mile to 7.5 and the gram. For this study, police officers will complete a supplemental Ohio crashes per road-mile to 6.5*. For the 7.5 accidents per accident report prcviding more information on road conditions and road-mile experience, approximately 200 miles of atud-damaged and pre-crash vehicle dynamics than is currently collected. The sup- 250 miles of undamaged pavement would be required to detect a 5% plemental form will cover such matters as pattern and depth of difference in per mile accident rates. For the 6.5 accidents per pavement wear at the accident site depth and pattern of water road-mile figure, 75 damaged and 115 undamaged road-miles will be coverage, severity of skidding patterns, and the officer's esti- needed to detect a 15% difference". The number of accidents and mate of the contribution of these factors to the crash sequence; a sample form is presented in Figure 0-3. The base accident per road-mile data was derived from published New York reports and from the fliRT Ohio Turnpike accident data Accident data presently available does not cover such itema file. The 80% car/non-icy figure was based on the Ohio experience in detail. Details about the road surface are often not given, or of 75.8% experience informally adjusted to reflect the higher proportion of susesertime passenger car nileage which the raw data if present, are presented in broad categories such as wet, snow- does not reflect due to use of four and one-half years data, the covered or Icy, and dry for surface condition, and as defective or lest haif-year of which does not cover the peak sulscscr months of July, August, and esrly September. not-defective for pavement condition. The first categorimation

1These figures have be en rounded up to the nearest 25 and S mile of surface condition permits no distinction among the various units respectively. c-ia C-20 32

I.. Water Distribution levels of water accumulation which may range from damp to covered Extent Distribution with water of an eatenstve depth. Presence of road defects are [3 Dry [1 Even probably only recorded for conditions which represent severe i,TJcse- C] Damp rj Puddled diate hazards such as pot holes and wash outs • with more subtle C] Wet i/Il" or less depth [I channeled factors such as polishing and studded tire dams ge normally ig- C] covered 1/4" or greater nored. SimIlarly, vehicle defects are most likely recorded only C] Slush when they have an obvious relationship to the accident, as when defective tires are indicated when a blow-out precipitates a 2. Precipitation crash. Vehicle dynamics are often indicated by a diagram on the [I None - Duration accident report. Unfortunately, most of the diagram information [3 Sprinkles C] Started less than 15 sin, prior is not reduced to digital form; the conventional approach is to C] Drizzle [3 Started 19 to 60 mm. prior indicate loss-of-control as the accident cause usually when no C] Moderate [1 Started 60 in. or more prior other explanation, like 'speed too fast for conditions", can be C] Driving applied, while crash reporting has been improving, the present data still reflect a preoccupation with assigning legal fault to I. Temperature 5. Tire tread a driver with a consequent neglect of road factors and a tendency C] hO's or lower R.F. L.H. to assess a single cause of the accident. Consequently, while C] BlF or above L.P. H.P. being useful in studying overall patterns, the present data do not provide the depth of detail to study a specific problem. 5. Skidding The hi-level approach, which has been used in a number of Before After Path studies, including the companion MCMII P report onw inter driving [I None C] [] Straight C] Greater than 360' effects of studded tire use, attempts to remedy some of the pro- C] Some C] [I Less than 90' C] Clockwise blems with current crash reporting procedures. In the bi-level C] Pronounced [] C] 9D - 180' C] Counterclockwise study, police offic ers complete a supplemental accident report C] lao' - 360 C] Fishtail form yielding detailed information about a specified class of accidents. The collected information is then merged with the ori- Degree of Contribution Degree of Co nridence ginal accident data base, and the appropriate statistical analysis [] Sole Cause of Crash [] Absolutely certain is conducted to test specific hypotheses about the problem being [] Major Factor [1 Highly confident studied. Useful information often can be gained through thehecir- cir- C] Contributing Factor C] Think probable culationculaticn of written instructions and forms to a wide number of [] Minor Factor C] Believe possible police agencies. However, information of the highest quality can [] No Tnfluence C] Pure speculation only be collected by working with a limits d number of agencies, C] Mitigating Factor by carefully inatructing field forces in the use of the form, and by continued field liaison with officers collecting the data. Figure C-3 Since the research plsn recocenends the use of a turnpike system Supplemental Accident Coding Form which is normally patrolled by a relatively small, permanently C-21 c-fl assigned state police ittachaunt, the carefully controlled ap- them by default. Particular items are discussed below. proach is easily achievable and should be followed. The water depth and pattern items seek to determine if stud- Most of the muggested form Items sic reasonably sel f-expla- damaged sections expertence substantial water depths more fre- natory and stem directly from the study effortc. The confi'Icnsce quently and if the part:cular water accumu lotion pattern can have and import ance sc ales for certain items are relatIvely unique. The an effect on out-of-control events. Water patterns are scfr,esut I- confidence scale has been assigned for items for which the officer cally outlined on the form. The water depth classifications have may not be able to make a direct observation but must rely on been chosen somewhat arbitrarily; indirect references from physical evidence,, witness reports, or Damp describes conditions when water is microscopically drivers' stateeents. The importance scale will aid in assessing present as evidenced by pavement color and tacti lesensa- the relative effect of pavement factors; it would differentiate tions of moistness, but with no discernible water cover. among these three hypothetical crashes; wet implies a diacernib te water cover less than one-quar- i. while traveling on a wet pavement, a driver, without ter inch deep. This range is most often associated with making any control movements, unexpectedly loses control slippery conditions, but dynamic hydroplaning can occur and the vehicle strikes some fixed object or another under eatreme conditions. vehicle. One Quarter to One Half Inch covers the range for which 2. while traveling on wet pavement, the driver makes a dynamic hydroplaning can happen under not unusual cir- violent evasive or braking aaneuver to avoid striking a cumstances, and for which slipperiness is almost always deer, 105 es control, and his vehicle impacts another present. vehicle or fixed object. Over One Half Tnch presents circumstances under which dy- ]. While traveling on wet pavement, the vehicle is struck namic hydroplaning can comceonly occur, and for which at an angle by another vehicle, goes out of control, and other hydroplaning forces, like differential drag among comes to rest without another impsct. wheels, can play an important role.

In the first crash, skid-producing conditions can be rated the The skid path items are included to determine, if stud-induced "most important" causative factor. In the second example, the dmsage patterns are ascocisted with more severe loss-of-control investigator would have to judge whether the skid would have oc- situations, which implies a greater inherent danger, since the curred under normal conditions. If so, wet pavement factors pro- likelihood of recovery is smaller for 5 more severe skid. Similar- bably can be considered "contributory", and if not, pavement ly, the pre-crash-saneuvers data is sought to determine if unexpec- factors would be rated 'very important', but not 'most important" ted loss-of-control is more frequently experienced on stud-dmaaged since the driver's actions precipitated the event. For the third, pavement. Traveling speed prior to crash and tire tread depth pavement factors can be judged "not important". The items on the information are requested since these elements frequently affect scale require subjective judgexents, but asking these opinions the incidence of hydroplaning. explicit will provide more information than excluding possibili- Pavement wear information is recorded only in a qualitative ties, either through arbitrary coding conventions or by omitting fashion, since it will be impractical to measure wear at each

c-23 c-24 33

accident site with the road damage survey techniques. The road condition, and skid characteristics. The accumulated data will damage survey results will be used as the primary classification be matched with the standard accident report information and the for pavement condition at the accident site; the milepost loca- road damage survey results, in order to assess significant dif- tion will be used to establish the necessary link. The officers ferences in accident patterns over various wear characteristics. report of pavement conditions will be used to insure against misclassification, either at the boundaries of pavement segments c.4 TRAFFIC FLOW MEASUSSNENT or for anomalous pavement wear in a particular section. All of the suggested research ao far has been directed to- The amount of data required to obtain statistical accuracy ward measuring the final effect of studded tire damage on acci- is difficult to predict a priori, since little information is dents. With this problem, as with many others, leolating the available on the distributions of the various factors. However, effect of s particular factor from the multiplicity of other fac- If 300 crashes are investigated for each wear condition, reasona- tors is quite difficult. Studies suggest one useful intermediate bly tight confidence intervals can be established for even rela- measure of effectiveness, changes in traffic flow associated tively infrequent cells. The 600 total accidents correspond to with damaged pavement sections. one year's wet weather crash experience on the Ohio Turnpike, To measure changes intraffic flow patterns, it is recom- and to a sosewhat shorter period on the New York Thruway. To mended that a time lapse photography technique be employed. The simplify administration, all wet weather accidents for a period position of a particular car would be photographed at several should be included; this will reduce the risk of bias in allowing intervals for a predetermined length of time. Using photometric the police to select accidents to meet a qucta. techniques, the relative position of the vehicle can be calcula- Data analysis will proceed like the comparison of the related. Aft er recor ding the information for a large number of vehi- tive frequency of wet weather accidents on damaged and undamaged cles, approximately 200, under a particular set of conditions, sections. As indicated earlier, detailed accident investigation the observations would be repeated at another site or sites, and should follow the accident data analysis, but it is possible to the results would be compared using standard analysis of variance perform this task in parallel with or prior to the previous task. techniques. However, the second alternative increases the risk of expending Traffic engineers have long felt that irregularities in traf- substantial effort without foreseesb le results. fic flow have contributed to accidents, with considerable recent In short, detailed accident investigation logically follows interest being generated by the conflict studies of Perkins and accident data analysis. The investigation will collect supple- others. The measurements have had a predominant empirical base mentary information on all wet condition crashes for a specified with strong statistical asaociati on shown between the variations oeriod. Police officers will collect the data in the course of in the traffic stream and the number of crashes occurring. their normal accident investigstiona, but to insure high quality, Theoretical explanations have been sprse, particularly in rela- constant field liaison must be maintained between the research ting driver actions in conflict situations to crashes. Despite group and the police. About 600 crashes will be needed. Informa- the lack of a good theory, it seems reasonable a aisji. that the tion will include items on wear patterns, water accumulation, tire greater the variation in vehicle paths, the more likely that they will either collid e with one another or leave the roadway and For example, for 300 observations, the 95% confidence interval about a base proportion of .10 is approximately ± .055. strike an object. ExamInation of Table 0-3, which was constructed

c-IS c-2a under extreme simplifying assumptions, indicates that the probabi- The reconended experimental program was directed toward lity of a crash may be substantially increased by slight increases measuring such fine changes in vehicle paths. An explicit con- in the standard deviations (or variance) of the vehicle path. For flict measurement app roach was considered, but it was not adopted, exaaple, an increase in the standard deviation of the path from since it was felt that conflicts, as usually defined, would occur 0.75 feet to 1.00 feet results in a 114-times increase in the pro- too infrequently under free- flow highway conditions. Instead the bability of leaving the road, and an increase in the standard de- approach used by the Wisconsin Highway Department in a reoent viation from 1.00 feet to 1.25 feet leads to a 31-times increase study was chosen as be ing more closely related to the underlying in the probability of two vehicles sideswiping each other. phenomena. The original study did not yield positive results, but this does not appear to be a consequence of the methodology. Table 0-3 Rather, the original study coiiected a limited amount of data Hypothetical Crash Probabilities as a Function which would have revealmd only gross differences, and observa- of Standard Deviation of vehicle Path1 tions were point measurements which are sosew hat less likely to reveal increased path varistions over some distance, particularly if vehicles tend to track each other*. The recocscended program Standard Probability2 Probability Deviation of Off-Road of aide-swipe3 increases the data to be collected and uses measurement of position over an extended distance to resolve these difficulties. .25 * The experiment should beconducted in two phases. The first .50 * * phase will involve testing the field equipment, determining the .75 0.32 x 1011 accuracy of measurements under particular conditions (nighttime 1.00 0.111 x l0_2 0,11 a l0' 1.25 0.82 x 10_2 0.311 a lD 2This is the probability that a vehicle's path will deviate by - more than 3' in only one direction. 1.50 - 0.23 5 10 0.23 x 10_2 1.75 0113 5 lo_l 0.76 x l0_2 3This is the probability, assuming independent changes in path, a 10_1 that separation between vehicles will be less than zero, given a 2,00 0.67 0.17 a lO_l mean lateral separation of six feet. 2.25 0.91 x lO_l 0.30 x 10 l01 *Thie second point is a rather fine one. If the variation of 2.50 0.11 0.119 a vehicle paths is truly random, then poir t measurements will yield 3.00 0.15 0.78 x 10 the same estimates of variance as will following the care over a 4.00 0.111 distance. If, though, the paths follow a cyclic pattern, there 0.23 is a possibility, perhaps remote, that the changes in frequency 5.00 0.27 0.20 and amplitude will be such that the app arent variation at a point remains unchanged. This possibility will be increased if the ve- 6.00 0.31 0.21 hicles tend to track each other, since at a particular point each vehicle will tend to maintain thesaae position in the roadway. 6 Further, since the observations were r.sde in an upstream direction Probability less than 0.5 x 10_ from overpaases, any path variation induced by stud damage might 1The tables are constructed assuming that vehicle or vehicles have have been obscured by the drivers' attempts tostmer e steady a sean path about the sidpoint of a twelve foot lane with a three course when confronted by the perceived constriction of the road- foot clearance between vehicle and edge of lane, and that the ir way from the overpass structure itself. actual position varies in a random fashion following the normal distribution- These are"instmntaneous" probabilities, in that they represent the probability that an event will involve a particular vehicle or vehicles at any instance in time. 0-25 0-27 34

and rain), and locating the mdx ir ,,r;i di:;t:so, over which photo- observations. This later expansion would significantly increase graphs can record vehie is mnas',n,'n I.:; cr11 ably. the cost of the data collecti on errort, hut would not double it, since eerie in costs associated with preparing a site for observa- At a maximum, 11111 uerlcs of' observations will be needed. This tions would remain fixed. represents observations at two different sites for each cornb.ln- tion of the conditions '.ndicatrd in Toule c-i, I f observations site selection will bearather straightforward process, the under nighttime or wet weather conditions prove to be impractical, principal problem being finding locations which satisfy the pave- the number of observations can be reduced to ii. A compromise ment wear/traffic volume requirements. The pavement wear infor- deaign, which neglecta wet weather effects, but which includes mation can be determined from the road damage survey information, nighttime effects of pavement marking wear, would require 48 total or if this experiment I. conducted independently, from qualitative site observations. As indicated earlier, 200 observations per asaeaament of the degree of damage along particular road segments. site would be sufficient to gcnorate hilt ctotlst icsl reliability*. The difficulty will be in locatine sites which have heavy damage but light traffic volume and vice versa. This problem might be simplified by using dafly fluctuations in traffic volume at a Table C-li particular site, rather than selecting sites with difrerant traf- Factors and Levels for Inclusion in Traffic Pattern Survey fic volumes. Yet, this procedure is somewhat more risky since it involves the oh ance of contamin sting the data with differences A. Pavement Wear P. Marking Wear particular to one site. The pavement marking levels can be easily N one ors light 1. None or alight acoderate 2. heavy established by conducting observations before and after the annual Heavy spring-time repainting. Lighting conditions pose no particular problem, other than the avoidance of dawn and dusk periods when C. Traffic Volume D. Light Condition glare will adversely effect driver performance. weather condi- Light 1. Day tions are the most problematical, and eliminating them from the Medium 2. Night Heavy design may avoid many scheduling probleas. Sites should be limited to non-interchange overpasses on straight, level sections to E. Weather condition;, avoid contamination from entering traffic, passing, and curve foi- clear iowing events. Hair, Many of the operational probleas will have to be solved during a pre-test period. Procedures suggested here can only be prelimi- Full experiment: AxhxCahlsg nary. F or each site, a survey will be required to determine ele- Minimum experiment; AssaC "ations and dietmnces from the point on the overpass whera the Compromise experiment; AaBsC + BsCxD(2). camera is to be mounted to a number of reference points along the roadway. Reference points should be fixed objects that are easily observations would be made of vehicles using the main driving visible and relatively close to the roadway, but should not be on lane, but the passing lane could be included by repeAting the the roadway itself, since it Is quite possible that they might be p = .99 obscured by other vehicles, dirt or debris, or obliterated by

c-3e

traffic wemr. Reference point should be at no more than 100 ssouid be sufficiently meg to I nsarc that the path of the vehicles intervals and should be found on both sidos of the roadway. between observation;; are Independent of one another. The two hour Standard delineator markings can be used for this purpose, or if period will be sufficiently long to insure that transient factors greater accuracy is destred, the survey crew can atake reference in the traffic stre am, .sucn an disablod vehicles on the shoulder, zi points at predetermined distances*. convoy of mobile home trans ports, or police patrol vehicles, will not grossly affect the cverall sample. At the same time, if the Photographing of the vehicles will be a relatively straight- two hour period is properly ct,omen, it will be sufficiently short forward process. Equipment required is cosceercially available to avoid the gross cycloc changes in traffic volume over the course and includes a )Secs motion picture camera equipped with a high- of the day. A five aecond filming period is recommended with quality wide-angle lena and variable-interval time lapse feature; frames being taken each 0.5 seconds. The five cecnmd period will an interval timer; a tape awitch traffic counter; an interval allow observation over approximately 500 feet (time' at 60 mph, 550' counter; and either a radio signal generator/receiver set or line at 75 mph), with approximately 50 feet between observations (44' transmiss ion equipment. The traffic counter can be installed in at 60 eph, 55' at 75 spa). the lane where traffic Is to be observed and would be connected to a digital interval counter. After passage of a determined After completing toe obsarvatiomc, the film will be processed number of vehicles, the interval counter would generate a signal and the observations translated into data on the vehicle path. to the camera control. The camera control would then activate The data reduction process can use available digitizing equipment the camera for the period necessary to take the desired number of which will record the co-ordinatoc of the surveyed reference frames. The camera should be tripod mounted and located on the points identified in the photograph and reference points on the downstream side of an overpass. A person should be stationed with vehicles. The vehicle reference points ideally would be the point the equipment at all tines to service it in came of malfunctaon of contact of the rear tires, but since, this may not always be oh- and to prevent theft and vandalism. servable some other points may need to be used. Three points are suggested: the lower left and right corners of the rear bumper, Specific values for several of the intervals mentioned in and the center of the license plate. Average values for the the preceding paragraph can be suggested. The basic observation height stove the pavement and the horizontal separation of these period should be two hours in length and should be tieed to avoid points can be used without introducing substantial error. The ob- peak traffic periods when vehicle paths are likely co be strcngly served coordinates of the reference points from the photographs and affected by saturation conditions. The number of vehicle passages the surveyed information shout the fixed reference points can then between observations would then be adjusted on the basis of traf- be processed through a relatively simple computer program to solve fic volume to achieve tne 200 vehicle observations within the two the necem mary trig onometric equations and to compute the variance hour period. The 36 second average Interval between observations of the vehicle path. The data can then be analyzed through stan- uard analysis of variance techniques to determine if significant This would simplify one portion of the data reduction process by allowing incorporation of standard coefficients for distance in differences are associated with stud-Induced damage factors. If computing the position In the lane. However, the convenience will the stud damaged areas t.how a significantiy hIgher variation in be minimal, since precise calculation of the vehicle will require site specific data on elevation of the camera, angle of declina- vehicle paths, than one might infer that such damage contributes tion of the camera, and grade and euperelevation of the road sur- to a factor which Is cossxonly thought to be associated with acci- face. dent causation.

C-3i c-a, 35

0.5 PRO6HAF COSTS professional level and anuld .1 nvolvt-a very direct data tranrfer The data gathering activities recommended are quite extensive operation. At a higher figure, obtalnOng the data might require and a presentation of lIkely program costs can be useful in jud- protracted negotiations with the supplying agency involving senior ging at what level each activity should be implemented. Such cc- personnel, and oncethe data were receIved, ovolution and super- timatea are presented in Tab act 0-5.0 through 0-5.4. As can be vision of a detailed cooputer and manual reco ding operation might sean in Table 0-5.0, the cost could range from $23,725 under very be needed. Costs were eased on approxie a te calary levels for the optimistic assumPtiona for the, most modest program, to $120,150 types of personoei involved at the university. Likewise, computer under very conservative assumptions for the most fully implemen- processing costs and equipment costs were derived from informal ted program. discussion with informed persona within the University. One major coat area, overh cad charges, has been omitted since rates and items Estimating costs for the program in advance in not an easy included vary radically from agency to agency. task, since several poorly known variablee will strongly affect the amounts involved. the costa will vary markedly with the a- Table 0-5.0 mount of data to be collected in the field and the type of person- Estimated Program Costs by 5'lajor Activity Area1 nel used to obtain it. In the case of data collection under the Activity Low !ilzii traffic flow measurement activity, itemized in Table 0-5.11, the Pavement Wear Survey $5,675 $25,500 possible variation is between $750 and $14,350. The low figure Accident Data Analysis $3,600 $111,900 was arrived at by assuming that 36 observations would be required, Detailed Accident that each observation could be conducted by a single student assis- Investigation $6,300 $36,100 tant, earning $3.50/hour, ins four hour period; and that only a Traffic Plow rgeaeureaent $7,750 $113,650 modest amount of local travel would be required. The high figure TOTAL $23,725 $120.150 of $1,350 assumed 1411 fietd observations, conducted over a six hour period, by two civil service rated engineering teohniclans earning $6.50/h our; in addition the higher figure reflected an Table C-5.1 assumed 50-mile travel distance for each observation and the neces- Estimated Program Costs for Pavement Wear Survey By Major sity of overnight trips for one-third of the observations. The other collection activities display simil ar variation for essen- Activity Low High tially the same reasona. Planning & Administration $1,500 $6,000 Other program costs ahosa some variation, but not as esten- Data Collection2 $1,675 $12,000 sively as data collection. The main source of variation in these Data Analysis $1,500 $7,500 areas arises primarily due to differences in optimietio and con- TOTAL $5,675 $25,500 aervative estimates of the amount of effort needed to complete a specific planning, progracsciing, or analysis task. For example, aAll expenses in this table reflect direct costs only, and do the planning and administration item under accident data analysis, not make provision for agency overhead charges. They do attempt Table 0-5.2, ranges in coat from $1000 to $3000. For the low fi- to reflect, where possible, all incremental costs, even though income instances the mervioaa may be 'free" to the project, such gure, the task might be supervised by someone at the junior as police time on supplemental aoc ident reports. 0-33 2See Table 0-1 for complete detail.

0-34 fable C-5.2 - Estimated Program Cant:, for Accident Data Aoaly.ia

Activity Low High Planning & Administration $1,000 $3,000 Computer Programming $ 600 62,1100 Manual Encoding of $ 0 $3,000 Supplemental Tnforaatien - Computer Processing $11000 $2,500 Statistical Analysis $1,000 $4,000 TOTAL $3,00 $14,900

Table 0-5.3 Estimated Program Costs for Detailed Accident Investigation

Activity Low High Planning & Administration $2,000 $6,000 - Data Collection $1,250 $211,000 Clerical Processing $3,000 of Forms 11 250 Computer Progreessing $ 800 $3,600 Computer Processing $1,000 $2,500 Data Analysis $1,000 $3,000

TOTAL $6,300 $36,100

Table 0-5.4 Estimated Program Costs for Traffic Plow Meamurement Activity Low High Planning & Adainlstration $1,500 $6,000 Equipment & Supplies $2,800 $8,600 Data Collection 11 750 $111,350 Photo rnterpretacion $ 1100 $7,100 Computer Programcsing 8 800 $2,000 Computer Processing $ 500 $2,000 Data Analysis . $1,000 $3,000 TOTAL $7,750 $43,650

0-35

APPENDIX I. Tire Nydroplaning and Wet Skid--General and induced By Studded Tire Pavement Wear BIBLIOGRAPHY

Aiibert. B.J. 'Tires and Hydroplsning." Society of Automotive A bibliography of rererences pertinent to the relationships Engineers Automotive Engimeering Congreaa. Proceedings, Jan. between studded tires, pavement wear, and accidents is presented i-12. lISA, Detroit, Michigan. 1968. in this appendix. References are divided into fifteen distinct Allbert, BiT., and Walker, S.C. "Tyre to Wet Road Friction At categories, and are listed alphabetically within each division. Nigh Speeds.' Institution of Mechanical Engineers, Automobile Division. Proceedings 1965-66, Vol. 180, Fart 2A. Mo. The fifteen categories are 105-121. 1966. Allen, CV. • and Smithson, F. Specialized Road Surfaces For Tire Hydroplaning and Wet Skid--General and Induced By Traction Test Purposes, SAE 7201169, Society of Automotive Studded Tire Pavement Wear Engineers. 1972. Safety Aspects of Road Repair rlaintenanoe Activities Blevins, I.E., and Dahir, S.W.M. , cosps. Selected Annotated Splash and Spray References on Skid Resistance, North Carolina State University, Department of Civil Engineering, Highway Research Program, 36 Lateral Placeaent Shifting Caused By Marking Wear and p., October 1968. Worn Wheel Paths Brown, J.R. An Experiment Cosparing the Performance of Road- Transverse 5°orces and Steering Effects Resulting Prom stones Used as Chippinga In Roiled Asphalt; A 5111 Derby Ring Roughened Pavement and Worn Wheel Paths Road (i953-65), RRL Report LR 63, Road Research Laboratory, Orowthorne, England. 1967. Noise and Vibration Effects on Driver Fatigue Resulting Prom Roughened Pavement Clark, SE., ed. Mech anics of Pneumatic Tires, NBS Monograph 122, National Bureau of Standards. November 1971. T. EjeCted Studs Thrown Pros'. High Speed Vehicles Vehicle Componont Degradation Daughaday, H., and Balmer, G.G. A Theoretical Analysis of Wy- dropianing Phenomena, Highway Research Record 311, pp. 107-108. Pavement Material Properties and Studded Tire-Induced Highway Research Board. 1970. Wear Characteristics S. DeVinney, WE. Factors Affecting Tire Traction, SAE 67011611 Studded Tire-induced Pavement Friction Changes Society of Automotive Engineers. 19 57. Pavement Marking Practices, Material Properties, and Dillard, J.H. • and Maupin. OW., Jr. 'Use of a Sprinkle Treat- Studded Tire Wear ment to Provide Skid Resistant Pavements.' Association of As- Studded Tire Performance Characteristics phalt Paving Technologiata. Proceedings, Vol. 40, pp. 383-395. 1971. Tire Stud Design and Performance Characteristics Accident Studies Invoiving Skidding and Studded Tire Use Oaliaway. B.M. Effects of Pavement Surface Characteristics and Textures on Skid Resistance, TTI Research Report 138-4, Texas Miscellaneous Transportation Institute. March 1971. ii. Gengenbsch. W. "Experlsencal Investigation Into the Aquaplaning of Vehicle Tyres on Wet Roads." Automobil-Industrie, Vol. 12, No. 11, November 1967, pp. 74-79.

0-2

12 Oengenbach. W. "Expariroenteile unte:'suchunis von Neifon auf 23. Home, W.B. • and Joyr.er, U.T. Pneumatic Tire Hydroplanong and nasser Pahrbahn (Experiaental Investigation of Tires on Wet some Effects on Vehicle Performance, SAE-9700, Society of Auto- Roads)," Teil 1' "Mesogerate und Messverfahren (Measuring in- motive Engineers. 1965. struments and Methods) ," All, Vol. 70, No. 3, 1963, pp. 83-89 Teil 2: "Meseergebnisse (Test Results) ,° API, Vol. 70, No. 8, 280. Home, W.B. , Yager, T.J. , and Taylor, G.R. Review of causes 1968. pp. 288-293. Tell 3: eweitereMessergebnisae (Further and Alleviation of Law Tire Traction on Wet Runways, NASA TN-D- Test Results)," ATI, Vol. 70, No. 9, 1968, pp. 310-316. 41106, National Aeronautics and Space Administration. April 1968. 13 Ones, C.G. "Factors Influencing the Friction Between Tyre and Road Under Wet Conditions." Institution of Mechanical Nosking, J.R. "The Role of Aggregates in Providing Skid-Resis- ansineers . Automobile Division. Symposium on Control of Vehi tant Roads." Influence of Road Surface on Skidding Symposium. des During Braking and Cornering. Proceedings, June 11, Proceedings, October 10, 1968, England. 1968. 1963, London. 1963. Keen, Harry M. Desi,n for Safety, Highway Research Record 214, Oiles, 0.0. Some Recent Developments in Work on Skidding pp. 7-12, Highway Research aoard. 1968. Problems at the Road Research Laboratory. Highway Research Record 46, pp. 43-59, Wighwsy Research Board. 1963. Kuner, M.N. Unified Theory of Nubber and Tire Friction, Eng; Res. Bull. B-94, Pennsylvania State linivers ity. July 1966. Gillespie, T.D. "Pavement Surface Characteristics and Their Correlation with Skid Resistance," Report Number 12, Pennsyi- guesser, NW., and Ne,er, W.E. "Tontative Skid-Resistance Re- vania State University, Joint Road Friction Program. June quirements for Main Rural Highways. ° International Colloquium 1965. on the Interrelation of Skidding Resistance and Traffic Safety on Wet Roads. Proceedings, Berlin. June 1968. Oough, V.a., and Badger, D.W. "Tyres and Road Safety." Inter- national Road Federation, Fifth World Meeting. Proceedings, Lander, F.T.W. , and Williams, T. The Skidding Resistance of September 18-211, 19i6, London. 1956. Wet Runway Surfaces With Reference to Surface Texture and Tyre Conditions, RRL Report LN 184, Road Research Laboratory, Crow- l. Harris, A.J. Road Surface Testure and the Slipperiness of Wet thorne, England. 1968. fi2ai.. Highway Research Record 2i4, pp. 18-23, Highway Research Board. 1968. Ludema, K.C. "Road Surface Texture,' Materials Research and Standards, Vol. il, No. 10, October 1971, pp. 13-16. Hawken, N., and Kennard, A.H. "New Developments in the Skid Resistance of Road Surfaces." World Highway conference 6th. Ludema, K.C. "Tires and Roads," Mechanical Engineering, Vol. Proceedings, October 4-10, 1968, Montreal, Canada. 1,76. 92, No. 12, December 1970, pp. 8_1i;. Hofferberth, W. "Improved Orip of Tyree on Wet Roads." Inter- Ludema, NC., and Lee, C.S. influence of Road-Surface Teature national Colloquium on the interrelation of Skidding Resistance on Tire-Road Interface Traction Lieits, ASME Publication 69- and Traffic Safety on Wet Roads. Proceedings.B erlin. June Lub-20, American SocIety Mechanical Engineora. 1969. 1968 Mahone, D.C. eVariatson in Highway Slipperiness Characteristics 20, Holmes, XE. • and St-joe, R.D. "Tyre Forces as Functions of with Location." aserican Society for Testing and Matertalt, Cornering and Brakinc Slipon Wet Road Surfaces." Institution 65th Annual Meetlnr, Symposium on Skid Resistance. Proceedings, of Mechanical Engineers. Proceedings 1968-69, Vol. lb3, Part June 29. 1962. 1962. 3N. 1969. 3. Mahone, 0.0., and Runkie. S.N. Pavement Friction Heeds • Vir- Home, Walter B. Tire Nydropianing and Its Effects on Tire ginia Highway Research Council, Charlotte, eille. Oecember 1971. Traction, Highway Research Record 214, pp. 24-33, Highway Re- search Roard. 1968. Martinez, Si., Lewla, J.M. , and Stocker, A.J. A Study of Variables Associated with Wheel Spin-Down and Hydropianing, Home, W.B. , and Dreher, R.C. Phenos,na of Pneumatic Tire Texas Transportation Institute. .yanuary 1972. droplaning. NASA TN-o-2056, National Aeronautics and Space Ad- ministration. November 1963. Maycock, G. Experteents on Tyre Treed Patterns, NRL Report LR 122, Road Research Laboratory, Crowthorne • England. 1967.

Nay cock, G. "Studi 5,; on the 1k hid inc Re:: I:: taaceorra-.;,;rrrpr- Car Tyren on Wet Surfaces." Institution ni MesF,anioai Otogin- 5-3 eers, Automobile Olvsaion . Proceedings 1965-66, Vol, 180, Part 2A, No. 4, pp. 122-141. 1 9b6. - 5-4

38. Meados, 3K. The ]f'ff'e-ct or Type Construction on Braking, RRL Holmes, 9., See, C., and Williams, A . R. 'A ComI,ined apprrfice Report LB 22'f , Road Research r,aiorntory, Crowthorne • England to the Optimlsatfon of 'tyre and Pavement interciction. ' iseri- 3969. can Chemical Society, Symposium on Tread Wear and Traction. Proceedings. April 1971. 39 Meyer. W.E. ",Co., Bc-suits os Re:,earcl, 00 Skid Control." Fdddration cnterr,atisnaie dec Sociéti€s d'Ing4nieurs des Tech- Colley, 5.5. • Christensen, A.P. , and Howien, W.J. Factors niques do I'Autcsobi.e, International Autemcbile Technical Affecting Skid Resistance and Safety of Concrete Pavements, Congress. 10th. Proceedings, Scciety of Automotive Engineera HRB Special Report 101, pp. 80-99, Highway Research Board. of Japan, Inc., Pr- 189-203. 1964. 3969.

80. Meyer, W .E. 'Whit M-,ket Pavements SI ippery?" Society of Au- Keill,AN., Jr. Wet Traction of Tracticniced Tires, NBS Tech- tomotive Engineers, Mid-Tear Meeting, Proceedings, June 7-1l, nical Rote 566, Hatisnal Bureau of Standards. February 1971. 1ILI. Montreal, canada. 1971. 58. Obertop, D .E.F. "Decrease of Skid-Resisting Properties of Wet 41. Meyer, WE., and Kum:ner. H.W. Mechanism of Force Transmission Road Surfmccs at High Speed,." American Society for Testing Between Tire and React, SAE 890A, Society of Automotive Engin- and Materials, 65th Anouaa Meeting, Symoosium on Skid Resis- eers. 1962. tance. Proceedings. June 29. 1962. 1962.

82. Meyer, W.r' ..an,] Ku]csser, H.W. Pavement Friction and Tempera- 55. Rose, 3.6. • and Ledbetter, W .B. Su,ssary of Surface Factors ture gffeetc, HRTS Special Report lOs, pp. 87-55, Highway Re- Influencing the Friclion Properties of Concrete Pavements, aearch Board. 1966. Highway Research Record 357, pp. 53-63, Highway Research Board. 1971. - 413, Meyer. WE., and Sthrock, M.D. Tire Friction, A State-of-the Art Review • Report 5344/115 Boo 4135, Pennsylvania State Universi- 55. Ross, T.P. "The Cortribution of Stone Size and Orientation ty. April 1969. in Skid Resistance." Australian Road Research Board, 5th Conference. Proceedings. August 1970. liii Moore, B.P. 'A Review of SqueezeFilms, Wear, Elsevier Pu- hltshing Company, amsterdam. September 396E Sabey, B.S. Road Surface Texture and the change in Skidding Resistance with Speed, RRL Report Rusher 20, Road Research 85, Moore, D.F. A Study of Tire-Surface snteract ion for the case Laboratory, Crowthorne, England. 1966. of Rolling on a Wet Surface. CAL Report Rusher YD-1969-V-2, Cornell Aeronautical Laboratory. 1965. Sabey, B .S. "Surface Texture Requirements for High Speed Roads, Stmndards and Skidding in Accidents." International 16. Moore, D.P. 'A Theory of Viscous Hydroplaning,' Tot. J. Mech. colloquium on the In:errelation of Skidding ReiT6ance and Sci. , Vol. 9, 1967, pp. 797-810. Traffic Safety on Wet Roads. Proceedings, ueriln. June 1963. 47. Moore, D.E. 'Drainage Criteria for Runway Surface Roughness,' Sabey, B.E. "The Road Surface in Relation to Friction and Journalof the Royal Aeroncutiesi 5QCi! Vol. 69, May 1965, We ar of Tyrea,' Road Tar, Vol. 23, No. I, 1969. pp. 15-21. pp. 337-382. Sabey, R .E. Wet Road Skidding Resistance at High Speeds on 49 Moore, OF. Prcdictiono f Skid-Resistance Gradient and a Vane ty of Surfaces on Al, RRL Report LB 313, Road Research orainago Charactoristaca of Pavements, Highway Research Record Laboratory, Crowthorne, England. 1968. 131, pp. 181-203, Highway Research Board. 1966. Sabey, e.g., and Lupton, G.E. 'Friction on Wet Surfaces of in Moore, D .F. The Logical Design of Optimum Skid-Resistant Sur- Tire-Tread-Type vulcanismies," Rubber & Chemistry and Techno- faces, HRB Speciaa Report lOi, pp. 39_l4i, Highway Research iDgI.. Vol. 37, Oct-Nov. 1964, pp. 878-893. Board. 1969. Sabey, BE., Williams, T. , and Lupton, G.M. Factors Affecting 50. Mortimer, p.r. • and Ludema, K.C. the affects of Salts on Road the Friction of Tires on Wet Roads, SAE 700376, Society of Au- Drying Rate,, Tire Friction, and Invisible Wetness, Highway tomotive Engineers. 1970. Research Record 391, pp. 85-58, Highway Ssesearch Board. 1972.

0-5 u-a

63. icnu lse , A.H . 'Types of Surl'acel'e t sic a ,iThe r gf feet on Weller, P.S., and Maynar'l, D. P. Methods of Text-tiring New Con- Skidding H esimianc,c Under Wet Condi e. ona." International Co 1 40- crete Road Surfaces to Provide Adequate Skidding Resistance, quiumon the Tnteeesation of' Skidd in,_9g5i0aisnce and Traffic RRL Report LB 290, Road Receirrf, Laboratory, prowl: horn,, Safety on Wet Roads. Proceedings, Berlin .Juee 3963. England. 1970.

641. ichulce, K .H. , and Beckman, L. "Friction Properties of Pave- Weller, 0.5., and Maynard, D.P. the Influence of Materials ments at Different Speeds. - Americar, Society for Testing and and Mix Design on the Skid Resistance Value and Texture Depth Materiale - 65th Annual Meeting, Symposium on Skid Resistance. of Concrete, RRL Report LR 337, Road Research Laboratory, Proceedings. June 29, 1962. 1952. Crowthorne, England. 1970.

Segel, L. • Ludema, K.C. , and Dugoff, N.J. New Area- for Tire 77. Williams, AR., Rolsss, T. , and Less, G. Toward the Unified Performance Methods, Highway Safety Research Institdte, Univer- Tire and Pavement for the Reduction of Skidding sSty of Michigan, ktn Arbor. 1968. Ral~~, SAS 720162, Society of Automotive Engineers. 1972. Sherwood, W.C. 'The Role of Aggregate Tope in Pavement Slip- Williams, AR., and Lees, G. 'Topographical and Petrographi- periness,' Kentucky U-ivecs ity ingioeergin Esperiment Station cal Variation of Road Aggregates and the Wet Skidding Resis- Bulletin, Vol. 20, No. 2, Septesb ec 1965. pp. 42-85. tance of Tyres,' Quarterly Journal of Engineering Oeol,9gI,, Vol. 2, No. 3, February 20, 1970, pp. 217-236. Ssith, H.A. 'A Susoary of Al' NCHRP Research in the Realm of Pavement Skid Resistance.' Asercan Association of State High- Yandell, M.D. 'The Effect of Surface Geometry on the Lubrica- way Officials, 57th Annual Meeting,. Proceedings. December ted Sliding Friction and Polishing of Roadstones," Australian 6-10, 1971, MiamI Beach, Fla. 1971. Road Research, Vol. 3, No. 10, June 1969. pp. 50-68.

Smithson, P.O., and Herzehg, PH. Investigation of Tire-Road 00. Teager, R.W. • and tuttle, J.L. Testing and Anal ysis of Tire Traction Properties, SAE 710091, Society of Automotive Engin- Hydroplaning, SAE Preprint 7204471, Society of Automotive eers. 1971. Engineers. May, 1972. -

Staughton, G.C. The affect of Tread Pattern Depth on Skidding Moore, D.E. 'An Elastohydrodynamic Theory of Tire Skidding." Resistance, RRL Report LR 323, Road Research Laboratory, Crow- 12th International Automobile Technical Congress. Proceedings, thorne, England. 1970. Report 2-02. 1968.

TO. Staughton, 0.0., and Williams, T. Tyre Performance in Wet nomsndl, H. Some Friction Data on Microsiped Tires, Pennsyl- Surface Conditions, RRL Report LR 355, Road Research Labora- vania State University. March 1965. tory, Crowthorne, England. 1970.

Stephens, J.E., and Goetz • W.H. "Designing wine Bituminous Mistures for High Skid Resistance," Highway Research Board II. Safety Aspects of Road Repair Maintenance Activities ProceedIngs, Vol. 39, 1960, pp. 173-190.

Sutton, J.R.O. 'use of an Epoxy Resin/Synthetic Aggregate Sy- An Estimate of Additional Posts to Repair Highways Damaged by stem to Impart Skid Resistance to Selected Road Surfaces." Continued Use of Studded Tires, State of Wisconsin • Department Australian Road Research Board, 5th Conference. Proceedings. of Transportation, Divon of Highways. March 1971. August 1970. Goodwin, C . A. "Ways to Safeguard the Motorists in Areas of 73. Weiner, B. "Skidding Resistance and Traffic Safety on Wet Construction," Traffic Quarterly, Vol. 9, No. 41, October 1955, Roads." International Colloquium on the Interrelation of pp. 543-562. Skidding Resistance and Traffic Safety on Wet Roads. Frocee- Berlin. June 1963. Meoarry, T.P. "Checklist for Safety at Highway Workaites," constructor, November 1956. pp. 38-40. 744, Weller, D .E. A Review of the Low-Speed Skidding Resistance of a Muster of Concrete Roads Containing Various Aggregates, RRL Mcoarry, T.P. 'Worksite Safety,' American Road Builder, Vol. Report LR iii, Road Research Laboratory, Crowthorne, England 43, December 1966, pp. 8-9, 21. 1970. 38

67. M-unro, OR, and Hung, NW. Motor Vehiolc Safety at Temporary 98. Lane, W.R. "Shatter of Drop in Streams of Air," md. Construct ion or Rep.lr. Sites on Highways, Highway Safety Re- Chem. , Vol. 113, June 1951. on. 1331-17. search Instit ute, University of Michigan, Ann Arbor. December 1968. See also items 83, 89. 80. Shepard, P.O. Fins: Report - Highway Signing for Safety, Vir- ginia Highway Research Council, VRRC Report 70-R54, llune 1971. TV. tateral Placement Shifting 85. The Effects of Studdcd Tires, State of Minnesots, Minnesota Cauled by Marking Wear and Worn Wheel Paths Departaent of Highways. March 1971.

Sc. Winter Damage to Ro:d Paverents organization for Economic. Co- 99 eordlhn, N.H. Re port onRffr'to of Pavement Wear on Lateral operation and Development. Road Research Group, Paris, France. Vehicle Placement, State of Wisconsin, Department of transpor- May 1972. tation, Division of Highways. December 31, 1971. 100 Gordon, Donald A. "ExperImental Isolation of the Driver's Visual Input," Human Factors, Vol. 8, No. 2, April 1966, pp. UI. Splash and Spray 129-137. 101 Hubbell, J.S. 'and Taylor, W.0. Evaluation of Pavement Mar- 91. Ohapoux, E. "oe 1'Effet des oisposiiifs de Protection Dits king to Designate Direction of Travel and Degree of Safety, 'Bavettes' sur les Protections d'Eau des Vehisules Automobiles Highway Research Record 221, pp. 23-40. Highway Research (Concerning the Sf facts of Protective Devises Called "Fender- Board. 1968. Flaps' en the Projection of Water by Motor Vehicles)", Techni- cal Aspects of Road Safety, Vol. 32, December 1967. ppT. -2.30. u i J V. affet of Pavem ent dge Mark ing o w o-Lan e Rur R aBawyt in , e , p p.-7. High- w y r ra. 0 Forbes, 3W. A Study of Accident Hazard in Relation to Fen- ders and Mudguard' ror Motor Vehicles, Highway Traffic Safety 103. Smith, P., and Srhonfeld, R. Pavement Wear Due to Studded Center, Michigan State University. East Lansing. April 1969. Tires and the Economic consequencel in Ontario, Highway Re- search Record 331, pp. 54-79, Highway Research Board. 1970. Ones, D.C. A Note on the Problem of Splashing When Vehicles Travel on a Wet Road, RRL Reaearch Mote RN/3766/OSG, Road loi Seydel, V.U. Die Bretnflussung des Fahrverhmltens ue durch Research Laboratory, crowthorne, England. May 1960. Leit- ond Randlinien cut Fretlandstraben (The Inflnne of Middle and Edge Msr:ings Upon Driving Behavior), Report Nor 954. Rams, TO. • and Wray, G.A. Sup press ion of Water Spray on Wet Kleine Fachbuchreihe Band 7, Verkehrspsychologlschen Tnatitut. Roads, Stevens Institute of Technology, Hoboken, N.J. 1971. Arbeiten, ITS (Tnstitote of Trsffic Psychology, Papers, III). January 1959. Ean, 1.0., Wray, GA., and Koib, E.G. The Formation of Truck Spray on Wet Road,, Final Report 1431Gavideon Labora- 105. Williston, R.M. Effert or Pavement Edge Mark ings on Operator tory. April 1970. Behavior, HRB Bulletin 265, pp. b-27, Highway Research Board. Koessler, P. "Kotfiugeluntersuchungen (Investigations of Nud- 1950. guarda)," Deutsche /raftfahrtforschung end Strassenveraehra- technik, Heft 175, 1965. (Translated by H.R. Fassbender, The Motor Industry Research Association, Runeston, Warwickshire, V. Transverse Poroem and Steering Effects England, Translation 32/65.) Reaulting From Roughened Pavement and Worn Wheel Paths Maycock, G. The Problem of Water Thrown Up by Vehicles on Wet Roads, Ministry of Transport, RRL Report 4, Road Research Bergman, W. • Pox, S,A. , and Saihol, E. 'Oynamios of the Auto- Laboratory, Harmondaworth, England. 1966. mobile in a cornering Maneuver On and Off the Highway.' First International Conference on the Mechanics of Soil-vehicle Systems. Proceedings, Torino, Italy. June 1961.

6-9 b-3M

chieaa, A. • and Oberto • I.. "Vertical Onciilatinct Rehavior Gough, V.E. "Tyresano AIr Saspension." Adv0nce' in Automo- of Six Types of Europeln Mnto, Con ." red6s'ation Interns- bile Engineering, Pergamon Press. 3963. tionale des Socidtifs d'Tngén neurs des Techniques de l'Auto- mobile (FTSITA). International Automohtle Technical Con- 100. Gough, V.a., and Jones, RB., and Udsil, W.S. Radial Ply, gress, 9th. Proceedings. 1962. Rigid Breaker Tires, SAE Preprint 990A, Society of Automo- tive Engineers. 1969. Galloway, W.J. , Clara, WE., and Kerric)c, J.S. Highway Noise--Measurement, Simulation, and Miaed Reactions, NCHRP Marshall, M.D., Pottinger, MG. • and Gibmon. G.E. Hibbling-- Report 76, National cooperative Highway Research Program. The Force and Moment Beh avior of Tires Over Edges, S.F. Good- 19644. rich Research center. (Unpublished). Gordon, 0G., Galloway, W.J. • Kugler, BA. • and Nelson, O.L. Masla, a., Diana, G. • and oiordana, F. Sul Comportamento di Highway ,Noise--A Design Guide for Highway Engineers, RCNRP un Autoveicolo in 000vaiu Terreno Accidentata (On the Be- Report 117, National cooperative Highway Research Program. havior of a cornering Vehicle on a Rough Surface), Milano 1971. Politecnico Institoto de Meccanica e cosrozionet delle Mac- chine, Italy. 1969. Guide for the Evaluation of Human Exposure to whole-Body Vibration,ISO TO lOB/NOT (Secr.-17), International orga- McHenry, H.S. "An Analysis or the Dynamics of Automobilem nization of Standardieation. Technical Coem,ittee 108. During Simultaneous Cornering and Ride MotIons.' Institution December 1968 (Unpublished). of Mechanical Rg4p5ere. iympos Sum on Handling of Vchicles Under Emergency conditions, Automobile Division. Proceedings. Guignard, J.C. Response to Low-Freo,uency Vibration, South- London. Jmnoary 1969. ampton University, Tnmtitute of aound and Vibration Research, England. October 1968. bk, W. 'The Dynamics of Vehicle SkId Oeviation as Caused by Road conditiona." First Internstional Skid Prevention 123, Hanes, RN. Human Sensitivity to Whole-body Vibration in Conference, proceedings. August 1959. Urban Transportation Systems: A Literature Review, Report APL/JHU-TPR 0011, The John Hopkins University. May 1970. 313. Gough, V.E. • and Whitehall, S.C. "Universal Tire Test Machine." Federation' ote r nationale des Sch.tiniA s d'Ingdnieurs McFarland, Ross A. "H uman and Environmental Factors of doThliqusdLAn e eA).9 terna onal Automobile Safety." SAE Golden Anniversary Suesmer Meeting. Autmobi e Tech. Congress. proceedings. 1962. proceedings, June 12-17. 1955. Atlantic City, N.J. 1955. McFarland, Ross A. Huaan Factors in Highway Transport Safet • Harvard School of Public Health, Boston, Mass. VI. Noise and Vibration Effects on Driver Fatigue 9i Resolttny from Roughened Pavement Mills, C.H.G. Noise Emitted by coasting Vehicles. The Mo- tor Industry Research Association, Lindley, England. 1970. Allen, G. 'Human Roaction to Vibration,' Journal of Environ- mental Sciences, Vol. 14, No. 5, September 1971. pp. 10-15. 127- Mitechke, N., and Kirschke, L. Untormachungen (Iber das Sohwlngunrsverhalten, des Schwingempfinden end die Schwin- A Review of Road Trifftc Noise: The ,.'orking Group on Research gungsbeanmpruchung del Menschen in fabrenden PereonSFRIt- Into Road Traffic Noise, RRL Report LR 357, Road Research La- wagen (Tnves tigations on the Vibration Behsvior, the vi- boratory, crowthorne • England. 1970. bration Perception, and the Vibration Strain of Men In Travelling PassengerCars), Rraunschweig Tnstitute of Techno- Butkunas, A.A. Random Vibration Anal yais and Vthicle Develop- logy, Inatitute of Astomotive Engineering, Germany. 1970. ment. SAE 690109, Society of Automotive Engineers. January 1969, 128. Oliver, RI., and Whttehaad, J.P. A Survey of the Ride IT,Characteristics of Contemporary Motor Cars, MIRA Report 1966/ Callow, G.D. A Surveyor the Noise in Private Cars, Report Motor industry Research Association. September 1966. 1970/14, The Motor Industry Research Association. October 1970.

6-12 s'dI Rathe, E.J. "Uber den La,'m den Straosenvcrkehrs," Acustica, Schwartc • OR., and Dierstein, P.C. Safety of Metal-Studded Vol. 17, No. 268, :966. Tires on Bare Pavement Surfaces, State of Illinois, Illinois ovisionof Highways, SprIngfield. March 1967. Ride and Vibration Data Manual, SAE J6a, Society of Automotive Engineers, 1965. Stclpp, D. "Untersuchungen Dber die Wl6'kung von Spikes bei normalen Fahrbahnbedingungen auf schnee- undeisfreien Hoeenblith, WA. • Stevens, EN., at al. Handbook of Acoustic Strassen, " Deutsche Kraftfshrtrorachung Und Straaaenverkehr- Noise Control, Vol. II, Noise and Man, WADO Tech. Report stechnik, Heft 190, 1967. ?&p. 52-20l4, Wright Air Development Center, Wright-Patterson Air Force Base, Dayton, Ohio. 1953.

Shoenberger, P.R. "Human Response to Whole-Body Vibration, VIII. Vehicle Component Degradation Perceptual and Motor Skills, Vol. 34, 1972, pp. 127-160. 144. Beladorf, MR., and Rice, R.S. eTests Show That Degradation Taibott, W.R. "Noise," Highway User, March, 1972, pp. 8-11, of Steering, Suspension Systems Ooean't Always Affect Car 38. Safety," SAE Journal of Automotive Engineering, Vol. 79, No, 7, July 1971, pp. 20-24. Tufts College Institute for Applied Experimental Psychology. Handbook or Human Engineering Data. Tufts College Bookstore, 185. Bird, M.D. • Belsdorf, MR., and Rice, P.S. Effects of Medford. Massachusetts. 1951. Steering and Suspension Component Degradation on Automobile Stability and Control, Part I, Final Report, CAL VC-2578-K-1, VanDeusen, B.S. Human Response to Vehicle Vibration, SAE 2, 3, 4, 5, 6, Cornell Aeronautical Laboratory. January 680090. Society of Automotive Engineers, January 1968, 1971.

Weiner, P.M. "Experimental Study of the Airborne Noise Genera- 186. Craggs, A. 'The Assessment of Pave Test Track Loads Using ted by Passenger Automobile Tires,' Noise Control, Vol. 6, Random Vibration Analysis." Noise and Vibration, Part 3, No. 4, 1960. pp. 13-16. Advances in Automobile Engineering, (OH, Tidtury, ed,) Symposium, Advanced School of Automotive Bngineering, Crsnfield, England, Proceedings, July 19604, Pergamon Press, London, 1965. VII. ijeCted Studs Thrown from High Speed Vehicles 147, Fancher, P.S., Ervin, M.D., Drote, P. Mecidam, CC,, and Segel, L,timit HandlIng Performance as Influenced by Or- Burke, J.E. , and McKenzie, L.J. Some Tests of Studded Tires gJ94J.tion of Steering and Suspension Systems, Final Report, In Illinois, State of Illinois, Illinois Division of Highways Highway Safety Research Institute, University of Michigan, Springfield. 1966. Ann Arbor. November 1972,

Evaluation of Studded Tires, Canadian Government Specirica- Jaeckel, N.H., and Swanson, SR.Random Load Spectrum Test tions Board, Department of Defence Production, Ottawa. March To Determine Durability of Structural Components of Automo- 18, 1967. tive Vehicle,, FISITA Paper 3-02, Federation Internationale des Sociétiis d'Ingdnieurs des Techniques de 1'Autosobihe. Hoghin, Li. Damage to Roads by Studded Tyres, RRL Report May 1968. 208, Road Research Laboratory, Crowthorne, England. 1968. Lipmon, C. Anal yris of FaIlures, SAE 6901494, Society of .140. Jensen, PA. • and 4orfhage, G.R. Preliminary Studies of Automotive angineers. May 1969. Studded Tires on Highway Pavements, Highway Research Record 136. pp. 8-23, Highway Research Board. 1966. MerrIll,MS. Steering Oiagnoa ia--A Study of Degraded Com- ponents Affecting In-Use Vehicle Handling, Contract FH-ll- 6629, Clayton Mfg. Co. September 1969. 141. Kullberg, C., and Ohlsson, S. oubbaoe dEck undermikninger 1963-1964. (Studded Tires Investigation 1963-1964), Report Swanson • SN. cval.uating Component Fatigue Performance Under 35, National Swedish Road andTraffic Research Institute, Programsosed Random, and Programmed Constant Amplitude Loading, (Statens VEg-Och Trafikinstitut), Stockholm, Sweden. 1965 SAE Preprint 690058, Society of Automotive Engineers, January 1969,

0'43 0-14

Wells, EN., Pttzmaurtec, Jr., Cuillioms, CE., Kahn, Huhtala, M. Hactarenka 1dm vii kutus erasal One, raittipsahly- SR., and Williams, P.O. An Investigation of used Car Safety atelni 1''; L,oe,-adal [a auonitot tu verta i8evm tutktmus (The Standards--Safety Index: Finaj eogt, Tech. Report 553 wearl ng Efrect Of Studded Tires on Certain Aspha t Pawersents; (8 VolsU, Operations Recearch, Inc. September 1969. Comparative snvectlgations by Hoad Testing Machine), Vaotson Taknillinen Tutkimus]aitos, helsInki, Finland. 1967, Jonsen, PA., and Xorfirare, G.R. Studded Tires - Their Ef- Ix. Pavement Material Properties and Studded fect on Pavement Wear and ['orforsance Characteristics • In- Tire-Induced Wear Characteristics terim Report, Stitrs of Minnesota, Minnesota Department of Highways, Orrico or Natorialm, St. Paul. 196i. Andersmon 0., LSlja, 8., Rosengren, A., Astroe, T. • and Maysar • 11,0., and .iobumsc[ser, W. "MRglichkeiten our Heseiti- Orbom, B. Paveaen, Wear Due to Studded Tyres Measured In the guns von Spi kesschr,dnn mittel-, Ubercug ddnnschaichtiger Test Road Machine of the National Swedish Road Research In- BelSp,o (Methods of El lsanat:on of Damage by Studded Tires on stitute, Special Report 83A, National Swedish Road and Traffic Thin Cvorlays)," Strsaao and Autobshn, Vol. 22, No. 12, Research Institute (Statens Vig-Och Trad'ikinstitut) Stockholm, December 1971. pp. 546-590. Sweden. 1969. 164, Keyser, J.B. iffeot of Studded Tires on the Durability of 159. Andersaon, 0., and Lihja, H. Road Wear Due to Studded Tyres Road Surfacing.. Highway aeaoarch Record 331, pp. 41-53, Measurements of Hood Profiles in the Stockholm Area, Progress Highway Heamarch Board. 3970, Re ort, National Swedish Road and Traffic Research Institute Statens Vig-Och Trafikinstitut) Stockholm, Sweden. November 165. Key:ror, , T.H. 'Mis Design Crib on a for Wosr-Resie tant Bitu- 1969. ainoui- PsvemerLt Surfaces.' Highway Research Board, 51st Annual Moeting.,, Proeerdiogrr, January 17-21, 1972, Sip, 155. Andermmon, 0., and Lil.ja, B. Stud lea of Pavement Wear Caused Washington, D.C. 1972. By Studded Passenger Car Tyres on Straight Teat Tracks, The 'Broma Trsck", National Swedish Road and Traffic Research Keyser, J.H. 'Resistance of Vardous Types of Hituminous Institute (Staten, VEg-Och Trafikinstitut), Report 3A, Stock- Concrete and Cement Concreto to Wear by Studded Tires.' holm, Sweden. 1972. Highway Research Board, Seth Annual Meeting. Proceedinas, January, 1972, Sip. Washington, D.C. 1971, 156. Bruselius, ND. • Soveri, U. • and Thrumann-Moe • T. epiggdek kenes slitasje pa vegdekkcr (Studded Tire Wear on Road Sur- Krukar, M. , and Cook, J.C. The Effect of Various Tire Studs faces)," Forhandlenger ved Nordiak vegteknimk forbuda. 10. on Oifrerent Pavement Materials and Textures • Washington Kcngress, Nordisk vegteknisk forbuds, Oslo • 1969, pp. 193-188. State University. May 22, 1972.

157. Burnett, W.B. , and Kearney, E.J. Studded Tires--Skid Resis- Lee, A. • Page, TA,, and oeCsrrera, R. Rffects of Carbide tance and Pavement Damage, Highway Research Record 136, pp. Studded Tires on roadway Surfaces, Maryland State Roads Com- 29-Ic, Highway Research Board. 1966. mission, Bureau of Research, Brooklsndville, Md. July 1965.

158. Diakiw, N.W. "Photograemietric Measurement of Concrete md Measurement of Wheelpath Wear on Wisconsin Highways'. A Two- Aiphaltic Pavement Wear in Metropolitan Winnipeg." North Year Report, State of Win cons in, Department of Transporta- American Snow Conference of the American Public Works Associa- tion, Division of Highwlyo . Msy 1971. tion. Proceedings, Detroit, Michigan. April 1972. Peffekoven, W. Laboruntorauchungen Iber den Abrieb von As- 319. Frocmi, H.J. , and Corkill, J.T. An Evaluation of Surfaco phaltholoitcnunter api korea renverkehr (Laboratory Investiga- Course Mixes Designed to Resist Studded Tire Wear, Report tionsof the Abrssion of Asphllt Pavements under Studded HRi71, Ontario Departsent of Highways, Downsview, February Tires), Moninli ii ke/Shell-Laboratonium, Amsterdam, Nether- 1971. lands, 1969.

160, Gragger, P. "010 Wlrkung. von Auftausalcen end won Spd kes auf ill. Preus, O.K. Effects of Studded Tires on Pavements and Traf- Teer- and bitumengobundenepahrbahcsdeclson (The Influence of fic Safety in Minnesota, State of Minnesota • Minnesota Dept. Road Salts and St,,ds on Tar and Bituminous Surracing,$) of Highwsys, MinneapolIs, 1972. Association of Roa,i and Traffic Engineers or Lower Sasony, Seminar. Proceed irigo . 2i October 1970, Hannovor • Germany. 1970,

0-15 0-16 40

Reouirand, B. Uaure des revAtements par lea pneus A c bus Wehner, B. 'Spikrc.r.'ti'.'tund Criffigkcts (Studded Tires and (Wear of Pavement Surfaces Caused by Studded Tires), Labora- Friction Value)," Stra:c.o urti Autobahn, Vol. 22, No. 0, toire Central des Fonts at ChsussBes, Nancy, Prance. Decent- January 1971. pp. 5-10. ber 1970. Rosengren • A. An Investigation Concerning Studded Tyres, Wehner, B. "The Road in ((elation to TraffIc Requirements." with Special Reference to Pavement Wear. Based on Litterature XIII Congress, Pernonont Tnternattnnal Association of Road Congresses. Proceedtngs, Tokyo. 1967. (sic) Studies, Special Report SIB, National Swedish Road and Traffic Research Institute (Statens VIg-Och Trafikinstitut), Stockholm, Sweden. 1969. Whtte, 0.A. , and J.nktna, J.C. Test of Steel Studded Snow Tires, State of Oregon, Oregon State Highway Department. Rosenthal, P. Haselton, FR. • Bird, M.D., and Joseph, P.J. Salem. 1966. Evaluation of Studded Tires'. Performance Data and Psveent Wear Measurements • NCHRP Report 61, National Cooperative Zichner, G. "Auftausal a, und Spikes-Reifen, lhre Auswirkung Highway Research Program. 1969.' our des Verhalten 'eon Pshrbahndecken (Road Salt and Studded Tires, Their effec; on the Condition of Highway Surraces)," Rosenthal, P., and Joseph, P.J. Deve lopmento f Methods for Strassenbauindustrie, Vol. 2, Heft 7, February 15, 1969, Assessing Pavement Wear by Studded Tires, Cornell Aeronauti- pp. 14-20. cal Laboratory, Inc., Buffalo, New York, January 17, 1908. 167. Zichner • C. "Die Auswirkung von Auft susolcen und Spikes- Seuterey, B. Revdteteents de ChaussPes Degats Nivernaux-- Reifen our des Vernalt en von Fahrbahndecken (The Effects of Action des Pneus S Cious (Winter Pavement oamage--The Influ- 0,-icing Salts and Studded Tires on Aaphntt Pavement Wearing ence of Studded Tires), Laboratoire Central des Fonts at Surfaces)." International Winter Road Congress. Proceedings. Germany. January 1969. Chaus—se7a_,57partment des Chausades • Paris, France. May 12, 1970. 188. Zichner, G. 'Die Reanaprunchung der Stratsendecken is Winter all. Sohulce, EN., and Beckmann, L. "Untersuchungendber den und MDgilchkei ten zur Vermir:derung des erh6hten V,rschleisses Verschleisa an Strassenoberfllchen durch Winterreifen sit (The Stress on Pavement Surfaces in the Winter and'Ponaiaili- Spikes (Studies on Pavement Wear in the Winter Due to Studded ties for Reducing :he Increasing Wear and Tear)," Strasae Tires)," Strasse Brucke Tunnel, Part 1: vol. 21, No. iO, und Autobahn, No. 22, Jahrgang, Heft 1, January 1971. pp. October 1969, pp. 267-273. Part 2: Vol. 21, No. 12, Decem- 1-5. ber 1969, pp. 309-318. Smith. P. • and Schonfeld, B. Studies of Studded-Tire Damage See also items 83, 90, 103, 137, 138, 140. and Performance in Datarto--Winter 1969-70, Report RR165, Ontario Department of Highways, nownaview. August 1970. Speer, T.L. , and German, J.W. Laboratory Evabuataon of X. Studded Tire-Induced Pavement Friction Changes Pavement Damage Caused by Studded Tires, Salt and Abrasive Sand, Final Report M71-il. American Oil Company, Hapervilla, Research and Development Department, Illinois. May 27, 1971. 169. Myyppl, J.M.I. Survey of the Skid Resistance Properties of Asphalt Surfaoings, State institute for Technical Research, 180, Studded Tire Evaluation Jr. New Jersey, State of New Jersey, Helsinki, Pinland. 1969. Hew Jersey Department of Transportation, Division of Research )0, and Evaluation. January 1967. Larsen, H. • and Boyhus, A. Om Anvende ate ar Pigdaek (Advantages and Drawbacks of Studded itres), Rapport 9, 181. Tons, F., and Ooeta, B.C. "Effects of Studded Tires." 56th Bidet for Trafiksikkerhedmforakning, Copenhagen, Denmark. Annual Michigan Highway Conierence. ProCeedings, March 9-10 1971. 1971, pp. 101-120, Grand Rapids, Michigan. See also items 103, 137, 140, 159, 167, 168, 178, 180, 183. 102. Wehner, B. 5n2pyuchung von Strsssenoberflichen durch Win- terreirenmit Spik6i('IKfluence Universitstof Studded Winter Tires on Pavement Surfaces), Technische , Berlin, Germany. July 196a,

0-17

and Del lAd ro , P. Performance of St ud,i.'u XI. Pavement Nark,ia tract isea, Mater ltd Proc'er rica, 201 . Pulley, A. L . , Stuti,l,',. 71cc boor Tires on Ice, CAL Report VJ-S704-V- a, Cornoll Aeronautic:, I LISJTatorv, Inc., Buffalo, New York. March 0966. The Effect of Certain 191. Chaiken, B. Cosoaciaoa of the Fe,'formnnm arc) Erc'nomy .,f Nasouasen, RE., Sad Doriese, A .D. Hot-Es truded Thermnpiasti c II Ighwsy .Strl ping Mai.cr ala and Tire-Road InterfaCs Parser: ore on Force and Moment Peri'or- Conventional Paint Strflhlnr, Dureau n f fiii'lic Rnntis , Office mance, Report A-2526, Oenerol Motors Corporation, Proving of Research and Development, Whahlngton, D.I. February 0TSd Section, Milford, Mlrhlgan. Juiy i969. 1969. Report on ComparatIve Performance of Permá-T-Gripper Studded 192. Improved Street Utllazation Through Traffic Engiiioerinr., TireswithUnstudded and Standard Studded Tires, Damas and May 197.2, HRB Special Report 93, 234p. Highway Research Board, 1967. Smith, Ltd., Toror.to, Canada. 204. Schallamach, A. 'Skid R,'aistance and Directional Control," Ohmpter 6, See also items 83, 90, 103, 137, 139, 178. Mechanics of Fneuc.atic Tires, S.K. Clark, Ed. • NBS Monograph 122, U.S. Government Printing Office, Washing- ton, D.C. Roves her 1971.

XII. Studded Tire Performance Chnractcrbstina 205, Smith, EM., Ewenc, WE., and Clough, D.J. Effectiveness of Studded Tires, Highway Research Record 352, pp. 39-69, Highway Research aoard. 1971. 193, A Critlesi Analysis of tt .. Minnenota Oepartmon,. of Ii ghaays Report "The Effects of Studded Tires,' Tire Stud Manufictu- 206. Smith, R,W. , and Clough, D.J. EffectIveness of Tires Under Toronto, Institute. March 1971, Winter Driving Conditions, Damas and Smith, Ltd. • rers Canadm/ljaterloo University, Department of Management 196, Baum, C.S. The Tire Stud Controversy anda •.w_çoncet-- Sciences. 1972. The Perma-T-aripper, Peroanence Corporation. April 09, 3972. 207, Whitehurat, E.A. , and Easton, A.H. AnEvaluation of Studded Tire Performance, NatIonal Safety Council, Commit tee on Bericht Ober Wlnterrelfen-PrOfungen(Report on Testan of Winter Tires with studs), Automobile Club of Switzerland. Winter Driving Hazards, Stevens Point, Wisc. 1967. acme. 1963. 208. Winter TeatReport, Annual, Rational Safety Council, Commit- tee on Winter Driving Hazards, Stevens Point, Wiso, 1967, Bernstein, B. 'Post Office Teats Tire Studs," Fleet Owner • Vol. 64, March 1965, pp. 268-259. 1968, 1969-1970. Bird, cD, • and Haseaton, P.R. Evaluation of Studded Tire Performance, CAL Report 159, Cornell Aeronautical Laboratory, See also items 305, 137, 1112, 157, 1610, 173, 174, 178, 180, Inc., Buffalo, be. 'Pork. September 1967. 181, l9D,

Dohbins, J.E. A R,.aativo PerformanceComparison Between Townsend and Con ventional Carbide Studs in Selected Toe And Highway Test Modes, NATO Project 20-17-07, Nevada Automotive XIII, Tire Stud Design and Performance Characteriatics Test Center. June-July 1971. Effectiveness of Studded Tires, Damas and Smith ttd. • Toronto, Cantz, R.J. New Tiro Stud Oevelopments, Kennsmetal, Inc., 109. 1972, Canada. 1970. Kengrip Division, Latroie, Pennsylvania. 'Mehr Sicherheit durch Winterreifen mit Spikes (More Safety 200, Kullaerg, 0. , and Klhlgren, B. Ur:deridkninp, cv vintordEck och alirakydd Sr frikcicaynpunkt (InvestIgation of FrIction Pro- Through Winter Tires with Spikes)," Automobil Revue/Revue perties of Winter °yres and Anti-Skid Devices). Rat ionai Swedish Automobile, Ho, 52. 1963. pp. 17, 19, 25, Road and TraffIc Research Institute (Statens VISE-nrC Trafikin- atitut), Stockholm, Sweden. 1960. 211. Miller, R.P,, II. Principles of Winter Tire Studs, Highway Research Record 171, pp. 1-13, Highway Research Board, 1967.

0-19 0-20 41

Miller • W. F. • IT. The WI nt.er Tire Stud, 4iihwnv Reaparet Re- 22', Nankina, ED. • Mortan. RB. • Ashkar, B., and Tutt, P.R. cord 136. pp. 1-6, Highway Research Bnmr,I. 1066. The Degree of Ir.fl-sence of Certain Factors PertainIng to the Vehicle and the Pavcscnt on Traffic Accidents Under Wet Con- Reits, B. Have We Finally Found the RIght Design ror Ti r,s ditions, Texas Highway Department, Highway Uesign Division, Studs? SAE 699269, Soc iety of Automotive Engi noera . 1969. Ri6rch Section, Auatin. Septeeber 1970. 214 Scheuba, N. "Das Zusarisnenspiel von Spikes und Reifen," 2211. Highway Safety. Design a9Lsperat ions,' Wet Weather Parlor- Afl, Vol. 73, No. 1, January 1971, pp. 18-22. manse-Studded Tires - Lack of Uniform Traffic Laws, House of Representatives, Congresa, Ninety-Second, First SessIon, 215. Scheuba, N. "Die Beanapruchung der Fahrbahn durch Spike- Subcommittee on Investigatsons and Oversight, Washington, Reifen deiaroinen Abrollen (The Stress on the Road Surface D.C. 1971. by Studded Tires at Consistent Speed)," Afl, Vol. 76, No. 1, January 3972. pp. 9-13. 225. Knofiacher, H. "h1uits and Experience Obtained from Road Accident Studies." International Colloquium on the Inter- 216 Soveri, U. 'Studded Tyre--A Problem in Finland," Shell Bi- relation of Skiddi;p Resistance and Traffic Safety on Wet tumen Review, Vol. 31, July 1970, 4p. Roads. Proceeding;, Berlin. June 1968. 217 Springenschmid, C. • and Sommer, H. 'Einfliissc von Spikerel- 226, Malo, A.?., and Mika, H.S. 'Accident Analysis of an Urban fen and Geschwindigkeit auf den Verschleiaa von Straseenbeton Expressway System," Highway Accident Studies, HRB Bulletin (Influence of Stucded Tires and Speed on Pavement Wear)." 2110, pp. 33_43, Highway Research Board. I96. Strasseund Autobahn, Heft 18, 1971. pp. 3-19. McCullough, B.F., asd h'ankins, gD. SkidResistanca Guide- 218 "U.S. Steel Research,ra Develop Wire Insert for Different lines for Surf ace Improvements on TexlBT'hwaya , Highway Type of Winter Snow Tire," Highway Research Hews, No. 26, Research Record 131., pp. 20I4-217, Highway Research Board. Winier 1967, pp. 39_40. 1966. Hrmad , J. Be, also liens 99, 130, 169, 156, 165, 167. 168, 1714, 176 181, 190, 193, 1911, 195, 197, 198, 201, 293. January 1971.

Percronek, K. Safety Effectivene-s of Studded Tires, Final Reoort, CAL Report VJ-2915-V-2, Cornell Aeronautical Labora- XIV. Accident Studi.eo Involving Skidding tory, Inc., Transportation Research Department, Buffalo, and Studded Tire Use New York. September 1971. Freuc, C.K. 'Effects of Studded Tires on Pavements in A Second Study of Relationships Between Tire Tread Depth and Minneaota," 56th .nnual Road School. Proceedings, Purdue the Likelihood of Accident Involvements, Highway Safety Foundation. May 1972. University, Lafayette, Indiana. pp. 12-22. 1970. First snternationa: SkId Prevention Conference. Proceedings. Baldwin, D.M. Assembly and Use of Acc Charlottesville, Virgins'a. August 1959. search Record 376, Highway Reaearch Boaidentrd. 1971.Data, Highway Re- Risker, B.S. Use ;f Accider,t Data to Identify Wet-Pavement Beaton, J.L. Skid-Proof Paveeents • California Division of Locationei n Perla:,lvania, Hi;hway Research Record 376. pp. Highways, Sacrasento. August 1970. 18-20, Highway Hate arch Board. 1971. Beckman, I. 'The Influence of the Skid Resistance of Roads Saiey, S.A. "Surface Texture Requirecants for High Speed on the Proportion of Accidents Under Wet Conditions.' Interna- Roads, Standards and Skidding in Accidents." International tional Colloquium on the Interrelation of Skidding Resistance Colloquium on the Interrelation of Skidding Resistance and and Traffic Safety on Wet Roads. Proceedings, B erlin. one Traffic Safety on Wet Roads. Proceedings, Berlin. June 1968. 1968.

0-21 5-22

TrafficCrssh Rc,pert Procedus,re State of Ohio, Department 21111. Hutchinson, J.W .,a.sd Kennedy, T.W. Medians of Divided High- of Highway Safety, Columbus, Ohio. September 1970. ways-Frequency and Nature of Encroachments, Engineering Experiment Station Bulletin 487, Illinois University, College Wehner, B. 'SkIdding Resistance and Traffic Safety on Wet of Engineering., Urbana. June 1966. Roads--Research Rcsults and Conclusions." International Colloqutum on the tntcrrc,lalion of Sktddan Re a nd 2115. Joerea, S.F. The influence of RainfallIntensity on Traffic, 1,Trff98 ic Safety on Wet Roa ds. Preedng,Belin r. Junea Flow, (Technical Report 44, The Store Drainage Research Pro- . ject), Department of Environmental Engineering Science, The John Hopkins University, Baltimore, Marylmnd. September Wolch, J.A. "Weather Cond itions and Road Accidents.' ALI5- 1965. tralian Road Raaerrch Board, 5th Conference. Proceedings, Part Three, pp. 190-208.u A gust 1970. 246. Jones, ER., and Coolsley, N.E. Effect of Rain on Freeway Capacity, TTI Research Report 24-23, Texas Transportation Institute. August 1969. See also itees 16, 344, 89, 100, 181, 193, 197. 2147. Eats, A. "Toward a Methodology of Traffic Safety Measure- ment and Program EvaluatIon," Accident Analysis and Preview Vol,3, Pergseon Press. 1971. XV. Miscellaneous 2148. Lucas, J. "Measurement of Geometric Roughness and Drainage of Water from Road Surfaces." International Colloquium on A Policy on Geometric DesiCn of Rural Highways. 1965., The Interrelation of Skidding Reaistanse and Traffic Safety American Association' of State Highway Officials, Washington, on Wet Roads. Proceedings, Berlin. June 1968. D.C. 1966. 2119. Mahone, D.C., and Shelburne, T.E. Research on Use of Stud- Cox, DR., and Lewis, P.A.W. The Statistical Analysis of ded Tirea A Review of Selected Literature, Virginia Highway Series of Events, Methuenand Company, Ltd., London, England. Research Council, Charlottesville. February 1967. 1966. Middleton, W.E.K. VIsion Through the Atmosphere, Toronto Frenk, R.S. • Skaar, D.E. , and Tennant, J.A. • Driver's Visual University Press, Toronto, Canada. 1966. Range Detection, SAN 720142, Society of Automotive Engineers 1972. Mood, AM. • and Oraybiil, F.A. Introduction to the Theory of Statistics, (Second Edttion), McGraw Hill Book Coepany, Gallaway, SM., and Rose, J.G. Macro-Thsture, Friction, New York. 1963. Cross Slope and Wheel Track Deoression Mesa urements on 61 Typical Texas Righa'ay Paveeenti, 'I'll Research Report 138-2, Moore, D.F. 'The Measureeent of Surface Texture and Drainage Texas Transportation Institute. June 1970. Capacity of Pavements." International Colloquium on the Interrelation of Skidding Resistance and Traffic Safety on Gallaway, u.N. • Rose, 1.0., and Schiller, RN. • Jr. The Wet Roads. ProceedIngs, Berlin. June 1966. Relative Effects If Several Factors Affecting Rain Water Psp3jia- on Pavement Surfaces, Texas Transportation Instttute, New York State Thruway Authortty, Twenty-Second Annual Re- College Station/Kentucky Department of Highways, Frankfort. port, 1971, Albany. 1972. January 1972. Parking Dimensiona, 1971 Model Cars, Engineering Notes 711, p147, Goodwin, W.A. Pre-Evaluatton of Pavement Materials for Automobile Manufacturers Association, Engineering and Techni- Skidding Resistance --A Review of U.S. Techniques, HRB Special cal Department, Detroit, Michigan. 1971. Report 101, pp. 69-79. Highway Research Board. 1969. 255. Rumar, K. Visible Distances in Night Driving- Effects of 2143. Oraybili, F.A. Ar Introduction to Linear Statistical Models Various Reflestanse Qualities or an and Object, 44th Re- Volume I, McGraw Hill Book Company, New York. 1961. port, University of Uppaala, Department of Psychology, Uppsala, Sweden. July 1967.

0-23 15-24 42

Smith, H.A. 'Status of Skid Resistance Research." Amerlan Association of State Htghwa Orricials Convention. Procee- !tl i97i. Studded Tires An unpublished Rthlior,raphv • Highway Safety Research Inrorsation Center, UnIversity of Michigan. Ann Arbor. April 1969. Tire Guide- 1972, Bennett Gsrfteld Publication, Farmingdale, New York. 1972. World Cars, 1972, Automobile Club of Italy. 1972. Voder, E.J. , and Mi lhous, R.T. Comparison of oifferent Methods of Measurirg Pavement Condition--Interim Report, National Cooperative Highway Research Program. NCHRP Report 7. 1964.

045 THE TRANSPORTATION RESEARCH BOARD is an agency of the National Research Council, which serves the National Academy of Sciences and the National Academy of Engineering. The Board's purpose is to stimulate research concerning the nature and performance of transportation systems, to disseminate information that the research produces, and to encourage the application of appropriate research findings. The Board's program is carried out by more than 150 committees and task forces composed of more than 1,800 administrators, engineers, social scientists, and educators who serve without compensation. The program is supported by state transportation and highway departments, the U.S. Department of Transportation, and other organizations interested in the development of transportation. The Transportation Research Board operates within the Commission on Sociotech- nical Systems of the National Research Council. The Council was organized in 1916 at the request of President Woodrow Wilson as an agency of the National Academy of Sciences to enable the broad community of scientists and engineers to associate their efforts with those of the Academy membership. Members of the Council are appointed by the president of the Academy and are drawn from academic, industrial, and govcrn- mental organizations throughout the United Srates. The National Academy of Sciences was established by a congressional act of incorporation signed by President Abraham Lincoln on March 3, 1363, to further science and its use for the general welfare by bringing together the most qualified individuals to deal with scientific and technological problems of broad significance. It is a privatc, honorary organization of more than 1,000 scientists elected on the basis of outstanding contributions to knowledge and is supported by private and public funds. Under the terms of its congressional charter, the Academy is called upon to act as an official—yet independent—advisor to the federal government in any matter of science and technology, although it is not a government agency and its activities are not limited to those on behalf of the government. To share in the tasks of furthering science and engineering and of advising the federal government, the National Academy of Engineering was established on December 5, 1964, under the authority of the act of incorporation of the National Academy of Sciences. Its advisory activities are closely coordinated with those of the National Academy of Sciences, but it is independent and autonomous in its organization and election of members. TRANSPORTATION RESEARCH BOARD NON-PROFIT ORG. National Research Council U.S. POSTAGE 2101 Constitution Avenue, N.W. PAID Washington, D.C. 20418 WASHINGTON. D.C.

ADDRESS CORRECTION REQUESTED PERMIT NO. 42970